WO2024104365A1 - 一种设备测温方法及其相关设备 - Google Patents
一种设备测温方法及其相关设备 Download PDFInfo
- Publication number
- WO2024104365A1 WO2024104365A1 PCT/CN2023/131686 CN2023131686W WO2024104365A1 WO 2024104365 A1 WO2024104365 A1 WO 2024104365A1 CN 2023131686 W CN2023131686 W CN 2023131686W WO 2024104365 A1 WO2024104365 A1 WO 2024104365A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- target area
- target
- visible light
- thermal infrared
- image
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 146
- 238000009529 body temperature measurement Methods 0.000 title claims abstract description 95
- 230000002159 abnormal effect Effects 0.000 claims abstract description 80
- 230000008569 process Effects 0.000 claims abstract description 80
- 238000012545 processing Methods 0.000 claims description 119
- 230000015654 memory Effects 0.000 claims description 67
- 230000005856 abnormality Effects 0.000 claims description 53
- 238000004364 calculation method Methods 0.000 claims description 21
- 238000001931 thermography Methods 0.000 claims description 21
- 238000012937 correction Methods 0.000 claims description 20
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012634 optical imaging Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 description 97
- 238000010586 diagram Methods 0.000 description 43
- 238000013528 artificial neural network Methods 0.000 description 32
- 239000011159 matrix material Substances 0.000 description 29
- 239000013598 vector Substances 0.000 description 26
- 230000006870 function Effects 0.000 description 24
- 238000007689 inspection Methods 0.000 description 19
- 238000003062 neural network model Methods 0.000 description 17
- 238000013473 artificial intelligence Methods 0.000 description 15
- 230000011218 segmentation Effects 0.000 description 14
- 238000001514 detection method Methods 0.000 description 10
- 230000001537 neural effect Effects 0.000 description 10
- 238000013439 planning Methods 0.000 description 8
- 230000004913 activation Effects 0.000 description 7
- 238000013500 data storage Methods 0.000 description 7
- 238000005516 engineering process Methods 0.000 description 7
- 230000008439 repair process Effects 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 6
- 238000013135 deep learning Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000010801 machine learning Methods 0.000 description 5
- 230000003287 optical effect Effects 0.000 description 5
- 101100233916 Saccharomyces cerevisiae (strain ATCC 204508 / S288c) KAR5 gene Proteins 0.000 description 4
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000003993 interaction Effects 0.000 description 4
- 230000003449 preventive effect Effects 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 230000004927 fusion Effects 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000012546 transfer Methods 0.000 description 3
- 239000011521 glass Substances 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000010191 image analysis Methods 0.000 description 2
- 239000003999 initiator Substances 0.000 description 2
- 230000036961 partial effect Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 230000000717 retained effect Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 101001121408 Homo sapiens L-amino-acid oxidase Proteins 0.000 description 1
- 101000827703 Homo sapiens Polyphosphoinositide phosphatase Proteins 0.000 description 1
- 102100026388 L-amino-acid oxidase Human genes 0.000 description 1
- 102100023591 Polyphosphoinositide phosphatase Human genes 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000005538 encapsulation Methods 0.000 description 1
- 230000008570 general process Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000004148 unit process Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J5/00—Radiation pyrometry, e.g. infrared or optical thermometry
- G01J5/48—Thermography; Techniques using wholly visual means
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01K—MEASURING TEMPERATURE; MEASURING QUANTITY OF HEAT; THERMALLY-SENSITIVE ELEMENTS NOT OTHERWISE PROVIDED FOR
- G01K11/00—Measuring temperature based upon physical or chemical changes not covered by groups G01K3/00, G01K5/00, G01K7/00 or G01K9/00
Definitions
- the embodiments of the present application relate to the field of equipment detection technology, and in particular to a method for measuring equipment temperature and related equipment.
- Preventive detection of power equipment in substations is an important means of maintaining power equipment. Measuring the temperature of power equipment is an important part of preventive detection. By measuring the temperature of power equipment, the normal temperature area and the abnormal temperature area in the power equipment can be determined, so as to judge whether the power equipment can work normally, that is, whether there is a fault in the power equipment.
- robots in substations can take thermal infrared images of a certain area containing target power equipment and send the images to a remote cloud server for image analysis. Since the image contains the temperature of each point in the area, the cloud server can determine the normal temperature area and abnormal temperature area of the target power equipment based on the image, and use this as the temperature measurement result of the target power equipment and promptly notify the substation staff so that the staff can inspect and repair the abnormal temperature area of the target power equipment.
- the power equipment in the substation is densely populated, and the area may contain not only the target power equipment but also the rest of the power equipment.
- the cloud server analyzes the thermal infrared image of the area, it is easily affected by the rest of the power equipment and mistakes the temperature of certain points near the edge of the rest of the power equipment as the temperature of certain points in the target power equipment. It misjudges the abnormal temperature area of the target power equipment, resulting in low accuracy of the temperature measurement results for the target power equipment.
- the embodiments of the present application provide a device temperature measurement method and related devices, which can accurately determine the temperature abnormality area of the target device without causing misjudgment, thereby effectively improving the accuracy of the temperature measurement results of the target device.
- a first aspect of an embodiment of the present application provides a device temperature measurement method, the method comprising:
- the staff issues instructions to the robot, and based on the instructions, the robot can determine the inspection task to be completed, that is, to complete the temperature measurement of the target device. Then, during the stage of performing the inspection task, the robot can control the camera to shoot the target area where the target device is located, thereby obtaining a visible light image of the target area and a thermal infrared image of the target area. It can be understood that the visible light image of the target area presents the entire target area, and the thermal infrared image of the target area presents the entire target area.
- the target area usually refers to an area that is demarcated to a certain range, which contains the target device and the regional device.
- the robot can send the visible light image and the thermal infrared image of the target area to the cloud server, and the cloud server can remove the visible light image and the thermal infrared image of the remaining areas of the target area except the target device from the visible light image and the thermal infrared image of the target area, thereby obtaining the visible light image and the thermal infrared image of the target device.
- the visible light image of the target device only presents the target device
- the thermal infrared image of the target device only presents the target device.
- the cloud server can determine the visible light image and the thermal infrared image of the temperature abnormality region in the target device from the visible light image and the thermal infrared image of the target device. It is understandable that the visible light image of the temperature abnormality region in the target device only presents the temperature abnormality region, and the thermal infrared image of the temperature abnormality region only presents the temperature abnormality region.
- the cloud server can directly use the visible light image and the thermal infrared image of the temperature abnormality area as the temperature measurement result of the temperature abnormality area of the target device and feed it back to the staff, so that the staff can find the temperature abnormality area of the target device based on the temperature measurement result and inspect the temperature abnormality area of the target device.
- the robot can control the camera to shoot the target area including the target device and other devices, obtain the visible light image and thermal infrared image of the target area, and send them to the cloud server.
- the server can process the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server can continue to process the visible light image of the target device and the thermal infrared image of the target device to obtain the thermal infrared image of the temperature abnormal area in the target device.
- the cloud server can determine the temperature measurement result of the temperature abnormal area of the target device based on the thermal infrared image of the temperature abnormal area, and report it to the staff so that the staff can repair the temperature abnormal area of the target device.
- the cloud server can remove the visible light image and thermal infrared image of the remaining devices from them to obtain the visible light image and thermal infrared image of the target device.
- the cloud server Since the visible light image and thermal infrared image of the target device only present the target device, in the process of analyzing these images, the cloud server will not be affected by the remaining devices, and accurately confirm the visible light image and thermal infrared image of the temperature abnormal area in the target device, so as to serve as the temperature measurement result of the temperature abnormal area of the target device. It can be seen that the cloud server accurately determines the temperature abnormal area of the target device without misjudgment, thereby effectively improving the accuracy of the temperature measurement result of the target device.
- the method further includes: calculating the thermal infrared image of the target area and the visible light image of the target area to obtain a depth image of the target area; processing the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device includes: processing the depth image of the target area, the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server may perform a series of calculations on the visible light image of the target area and the thermal infrared image of the target area to obtain the depth image of the target area. After obtaining the depth image of the target area, the cloud server may use the depth image of the target area to remove the visible light images of the remaining areas of the target area except the target device and the thermal infrared images of the remaining areas from the visible light image of the target area and the thermal infrared image of the target area, thereby obtaining the visible light image of the target device and the thermal infrared image of the target device.
- the method further includes: aligning the visible light image of the target area to the thermal infrared image of the target area to obtain an aligned visible light image of the target area; calculating the thermal infrared image of the target area and the visible light image of the target area to obtain a depth image of the target area includes: calculating the thermal infrared image of the target area and the aligned visible light image to obtain a depth image of the target area; processing the visible light image of the target area and the thermal infrared image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device includes: processing the aligned visible light image of the target area, the thermal infrared image of the target area, and the depth image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- the cloud server may align the visible light image of the target area with the thermal infrared image of the target area based on the first model (including coarse correspondence and fine alignment), thereby obtaining the aligned visible light image of the target area.
- the cloud server may also perform a series of calculations on the aligned visible light image of the target area and the thermal infrared image of the target area based on the second model, thereby obtaining a depth image of the target area.
- the cloud server can complete instance segmentation of the aligned visible light image of the target area, the thermal infrared image of the target area and the depth image of the target area based on the third model and the fourth model, thereby accurately obtaining the visible light image and the thermal infrared image of the target device.
- the method further includes: based on the depth image of the target area, determining the thermal infrared image of the foreground in the target area and the visible light image of the foreground in the thermal infrared image of the target area and the aligned visible light image, wherein the foreground includes the target device; processing the aligned visible light image of the target area, the thermal infrared image of the target area and the depth image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device includes: processing the depth image of the target area, the thermal infrared image of the foreground and the visible light image of the foreground to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server may also, based on the depth image of the target area, remove the visible light area of the background in the target area and the thermal infrared image of the background in the target area from the thermal infrared image of the target area and the aligned visible light image of the target area, retain the thermal infrared image of the foreground in the target area and the visible light image of the foreground in the target area, wherein the visible light image of the foreground in the target area only presents the foreground of the target area, and the thermal infrared image of the foreground in the target area only presents the foreground of the target area, wherein the foreground of the target area includes the target device and the remaining devices, and the background of the target area is the environment where the target device is located.
- the cloud server After obtaining the thermal infrared image of the foreground in the target area and the visible light image of the foreground in the target area, the cloud server can complete instance segmentation of the thermal infrared image of the foreground in the target area, the visible light image of the foreground in the target area and the depth image of the target area based on the third model and the fourth model, thereby accurately obtaining the visible light image and the thermal infrared image of the target device.
- the visible light image and the thermal infrared image of the target area are processed to obtain the target area.
- the visible light image of the device and the thermal infrared image of the target device include: segmenting the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target sub-area and the thermal infrared image of the target sub-area, the target sub-area being the area occupied by the target device and a part of the remaining devices in the target area; performing secondary segmentation on the visible light image of the target sub-area and the thermal infrared image of the target sub-area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the instance segmentation includes the first segmentation and the second segmentation.
- the cloud server can perform the first segmentation on the thermal infrared image of the target area and the visible light image of the target area based on the third model and the depth image of the target area, thereby obtaining the thermal infrared image of the target sub-area and the visible light image of the target sub-area.
- the cloud server can also perform secondary segmentation on the visible light image of the target sub-area and the thermal infrared image of the target sub-area based on the fourth model, thereby accurately obtaining the visible light image of the target device and the thermal infrared image of the target device.
- the method further includes: obtaining the distance between the camera and the temperature abnormal area based on the visible light image of the target area and the thermal infrared image of the target area; adjusting the temperature measurement result based on the distance between the camera and the temperature abnormal area and the preset corresponding relationship to obtain the adjusted temperature measurement result of the temperature abnormal area, and the preset corresponding relationship is used to indicate the corresponding relationship between the distance and the temperature correction value.
- the cloud server is also provided with a preset corresponding relationship, which is used to indicate the corresponding relationship between the distance and the temperature correction value.
- the cloud server can determine the temperature correction value corresponding to the distance between the thermal imaging camera and the target device based on the preset corresponding relationship and the distance between the thermal imaging camera and the target device (the distance can be obtained from the depth image of the target area), and then adjust the thermal infrared image of the temperature abnormal area based on the temperature correction value to obtain the adjusted thermal infrared image of the temperature abnormal area.
- the cloud server can use the visible light image of the temperature abnormal area and the adjusted thermal infrared image of the temperature abnormal area as the adjusted temperature measurement result of the temperature abnormal area in the target device, and feedback it to the staff.
- the cloud server can make the adjusted thermal infrared image of the temperature abnormal area in the target device closer to the actual temperature of the temperature abnormal area in the target device, which is conducive to improving the accuracy of the temperature measurement results of the target device.
- the cloud server can automatically obtain the distance between the camera and the temperature abnormal area in the target device based on the depth image of the target area, so as to determine the temperature correction value corresponding to the distance, and use the temperature correction value to complete the adjustment of the temperature measurement results. It can be seen that the entire process of temperature correction can be automatically completed by the cloud server, without the need for staff to operate, which can reduce the cost of manual operation.
- controlling a camera to photograph a target area to obtain a visible light image of the target area and a thermal infrared image of the target area includes: controlling a camera to photograph the target area at a preset position and at a preset angle to obtain a visible light image of the target area and a thermal infrared image of the target area, the distance between the camera at the position and the target device in the target area is within a preset range, and when the camera photographs the target area at the angle, the degree of overlap between the target device and the remaining devices is less than a preset threshold.
- the cloud server in the inspection planning stage, can select a preset position and a preset angle for the robot based on the images of the target area photographed by the robot at different positions and at different angles according to certain conditions, and these conditions include: when the robot is at the position, the distance between the camera and the target device in the target area is within a preset range, and when the robot controls the camera to photograph the target area at the angle, the degree of overlap between the target device and the remaining devices is less than a preset threshold. It can be seen that the cloud server can automatically plan the optimal position and angle for the robot according to the aforementioned conditions, and the factors considered are relatively comprehensive. This allows the robot to capture the optimal image at the optimal position and angle during the inspection execution phase, which is not only efficient but also improves the accuracy of the temperature measurement results of the target equipment.
- the robot's camera includes an optical imaging camera and a thermal imaging camera.
- a second aspect of an embodiment of the present application provides a device temperature measurement device, which includes: a shooting module, which is used to control a camera to shoot a target area to obtain a visible light image of the target area and a thermal infrared image of the target area, wherein the target area includes a target device; a first processing module, which is used to process the visible light image of the target area and the thermal infrared image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device; a second processing module, which is used to process the visible light image of the target device and the thermal infrared image of the target device to obtain a thermal infrared image of a temperature abnormality area in the target device; and a determination module, which is used to determine the temperature measurement result of the temperature abnormality area based on the thermal infrared image of the temperature abnormality area.
- the robot can control the camera to shoot the target area including the target device and other devices, obtain the visible light image and the thermal infrared image of the target area, and send them to the cloud server.
- the cloud server can process the visible light image and the thermal infrared image of the target area to obtain the visible light image of the target device. and the thermal infrared image of the target device.
- the cloud server can continue to process the visible light image of the target device and the thermal infrared image of the target device to obtain the thermal infrared image of the temperature abnormal area in the target device.
- the cloud server can determine the temperature measurement result of the temperature abnormal area of the target device based on the thermal infrared image of the temperature abnormal area, and report it to the staff so that the staff can repair the temperature abnormal area of the target device.
- the cloud server can remove the visible light image and thermal infrared image of the remaining devices from them to obtain the visible light image and thermal infrared image of the target device.
- the cloud server Since the visible light image and thermal infrared image of the target device only present the target device, in the process of analyzing these images, the cloud server will not be affected by the remaining devices, and accurately confirm the visible light image and thermal infrared image of the temperature abnormal area in the target device, so as to serve as the temperature measurement result of the temperature abnormal area of the target device. It can be seen that the cloud server accurately determines the temperature abnormal area of the target device without misjudgment, thereby effectively improving the accuracy of the temperature measurement result of the target device.
- the device also includes: a calculation module, which is used to calculate the thermal infrared image of the target area and the visible light image of the target area to obtain a depth image of the target area; a first processing module, which is used to process the depth image of the target area, the visible light image of the target area and the thermal infrared image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- a calculation module which is used to calculate the thermal infrared image of the target area and the visible light image of the target area to obtain a depth image of the target area
- a first processing module which is used to process the depth image of the target area, the visible light image of the target area and the thermal infrared image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- the device also includes: a third processing module, used to align the visible light image of the target area to the thermal infrared image of the target area to obtain the aligned visible light image of the target area; a calculation module, used to calculate the thermal infrared image of the target area and the aligned visible light image to obtain a depth image of the target area; and a first processing module, used to process the aligned visible light image of the target area, the thermal infrared image of the target area, and the depth image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- a third processing module used to align the visible light image of the target area to the thermal infrared image of the target area to obtain the aligned visible light image of the target area
- a calculation module used to calculate the thermal infrared image of the target area and the aligned visible light image to obtain a depth image of the target area
- a first processing module used to process the aligned visible light image of the target
- the device also includes: a fourth processing module, which is used to determine, based on the depth image of the target area, the thermal infrared image of the target area and the aligned visible light image, a thermal infrared image of the foreground in the target area and a visible light image of the foreground, wherein the foreground includes a target device; and a first processing module, which is used to process the depth image of the target area, the thermal infrared image of the foreground and the visible light image of the foreground to obtain a visible light image of the target device and a thermal infrared image of the target device.
- a fourth processing module which is used to determine, based on the depth image of the target area, the thermal infrared image of the target area and the aligned visible light image, a thermal infrared image of the foreground in the target area and a visible light image of the foreground, wherein the foreground includes a target device
- a first processing module which is used to process the depth
- the first processing module is used to: segment the visible light image and the thermal infrared image of the target area to obtain a visible light image and a thermal infrared image of the target sub-area, where the target sub-area is the area occupied by the target device and a part of the remaining devices in the target area; perform secondary segmentation on the visible light image and the thermal infrared image of the target sub-area to obtain a visible light image and a thermal infrared image of the target device.
- the device also includes: an acquisition module, used to acquire the distance between the camera and the temperature abnormality area based on the visible light image of the target area and the thermal infrared image of the target area; an adjustment module, used to adjust the temperature measurement result based on the distance between the camera and the temperature abnormality area and a preset correspondence relationship to obtain an adjusted temperature measurement result of the temperature abnormality area, and the preset correspondence relationship is used to indicate the correspondence between the distance and the temperature correction value.
- an acquisition module used to acquire the distance between the camera and the temperature abnormality area based on the visible light image of the target area and the thermal infrared image of the target area
- an adjustment module used to adjust the temperature measurement result based on the distance between the camera and the temperature abnormality area and a preset correspondence relationship to obtain an adjusted temperature measurement result of the temperature abnormality area, and the preset correspondence relationship is used to indicate the correspondence between the distance and the temperature correction value.
- a shooting module is used to control a camera to shoot a target area at a preset position and at a preset angle to obtain a visible light image of the target area and a thermal infrared image of the target area.
- the distance between the camera at the position and the target device in the target area is within a preset range.
- the degree of overlap between the target device and other devices is less than a preset threshold.
- the camera includes an optical imaging camera and a thermal imaging camera.
- a third aspect of an embodiment of the present application provides a device temperature measurement apparatus, which includes a memory and a processor; the memory stores code, and the processor is configured to execute the code.
- the device temperature measurement apparatus executes the method described in the first aspect or any possible implementation method of the first aspect.
- a fourth aspect of an embodiment of the present application provides a circuit system, which includes a processing circuit, and the processing circuit is configured to execute the method described in the first aspect or any possible implementation manner of the first aspect.
- a fifth aspect of an embodiment of the present application provides a chip system, which includes a processor for calling a computer program or computer instructions stored in a memory so that the processor executes the method described in the first aspect or any possible implementation method of the first aspect.
- the processor is coupled to the memory through an interface.
- the chip system also includes a memory, in which a computer program or computer instructions are stored.
- a sixth aspect of the embodiments of the present application provides a computer storage medium storing a computer program.
- the program When the program is executed by a computer, the computer implements the method described in the first aspect or any possible implementation manner of the first aspect.
- a seventh aspect of the embodiments of the present application provides a computer program product, which stores instructions. When the instructions are executed by a computer, the computer implements the method described in the first aspect or any possible implementation method of the first aspect.
- the robot under the instruction of the staff, can control the camera to shoot the target area including the target device and the remaining devices, obtain the visible light image of the target area and the thermal infrared image of the target area, and send them to the cloud server.
- the cloud server can process the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server can continue to process the visible light image of the target device and the thermal infrared image of the target device to obtain the thermal infrared image of the temperature abnormal area in the target device.
- the cloud server can determine the temperature measurement result of the temperature abnormal area of the target device based on the thermal infrared image of the temperature abnormal area, and report it to the staff so that the staff can repair the temperature abnormal area of the target device.
- the cloud server can remove the visible light image and thermal infrared image of the remaining devices from them to obtain the visible light image and thermal infrared image of the target device.
- the cloud server Since the visible light image and thermal infrared image of the target device only present the target device, the cloud server will not be affected by other devices during the analysis of these images, and accurately confirm the visible light image and thermal infrared image of the temperature abnormal area in the target device as the temperature measurement result of the temperature abnormal area of the target device. It can be seen that the cloud server accurately determines the temperature abnormal area of the target device without misjudgment, thereby effectively improving the accuracy of the temperature measurement results of the target device.
- FIG1 is a schematic diagram of a structure of an artificial intelligence main framework
- FIG2a is a schematic diagram of a structure of a device temperature measurement system provided in an embodiment of the present application.
- FIG2b is another schematic diagram of the structure of the device temperature measurement system provided in an embodiment of the present application.
- FIG2c is a schematic diagram of a device related to device temperature measurement provided in an embodiment of the present application.
- FIG3 is a schematic diagram of the architecture of the system 100 provided in an embodiment of the present application.
- FIG4 is a schematic diagram of a flow chart of a device temperature measurement method provided in an embodiment of the present application.
- FIG5 is a schematic diagram of a structure of a calibration plate provided in an embodiment of the present application.
- FIG6a is a schematic diagram of a calibration process provided in an embodiment of the present application.
- FIG6b is another schematic diagram of the calibration process provided in an embodiment of the present application.
- FIG6c is another schematic diagram of the calibration process provided in an embodiment of the present application.
- FIG7 is a schematic diagram of refined matching provided in an embodiment of the present application.
- FIG8 is a schematic structural diagram of a first model provided in an embodiment of the present application.
- FIG9 is a schematic diagram of a structure of a second model provided in an embodiment of the present application.
- FIG10 is a schematic diagram of heterogeneous binocular disparity provided in an embodiment of the present application.
- FIG11 is a schematic structural diagram of a third model provided in an embodiment of the present application.
- FIG12 is a schematic structural diagram of a fourth model provided in an embodiment of the present application.
- FIG13 is a schematic structural diagram of a fifth model provided in an embodiment of the present application.
- FIG14 is a schematic diagram of a structure of a device temperature measuring device provided in an embodiment of the present application.
- FIG15 is a schematic diagram of a structure of an execution device provided in an embodiment of the present application.
- FIG16 is a schematic diagram of a structure of a training device provided in an embodiment of the present application.
- FIG. 17 is a schematic diagram of the structure of a chip provided in an embodiment of the present application.
- the embodiments of the present application provide a device temperature measurement method and related devices, which can accurately determine the temperature abnormality area of the target device without causing misjudgment, thereby effectively improving the accuracy of the temperature measurement results of the target device.
- Preventive detection of power equipment in substations is an important means of maintaining power equipment. Measuring the temperature of power equipment is an important part of preventive detection. By measuring the temperature of power equipment, the normal temperature area and the abnormal temperature area in the power equipment can be determined, so as to judge whether the power equipment can work normally, that is, whether there is a fault in the power equipment.
- robots in substations can take thermal infrared images of a certain area containing target power equipment and send the image to a remote cloud server for image analysis.
- the cloud server can determine the thermal infrared image of the target power equipment, that is, the temperature of all points of the target power equipment. Therefore, the cloud server can determine the normal temperature area and the abnormal temperature area of the target power equipment, and use this as the temperature measurement result of the target power equipment and promptly notify the substation staff so that the staff can inspect and repair the abnormal temperature area of the target power equipment.
- the power equipment in the substation is densely populated, and the area may contain not only the target power equipment but also the rest of the power equipment.
- the cloud server analyzes the thermal infrared image of the area, it is easily affected by the rest of the power equipment and mistakes the temperature of certain points near the edge of the rest of the power equipment as the temperature of certain points in the target power equipment. It misjudges the abnormal temperature area of the target power equipment, resulting in low accuracy of the temperature measurement results of the target power equipment.
- the robot often needs staff to set a certain position for the robot so that the robot can stop at the position and take pictures of the area containing the target power equipment.
- the factors considered in manually selecting the position are often relatively simple, resulting in the thermal infrared image of the area being not the optimal image, which is not only inefficient but also reduces the accuracy of the temperature measurement results of the target power equipment.
- the robot uses a thermal imaging camera to obtain thermal infrared images of the area.
- the temperature of any point in the image i.e., any pixel in the image
- the thermal imaging camera based on the received radiation energy. Since the radiation energy is affected by the distance between the target power equipment and the camera, the measured temperature of each point on the target power equipment will deviate from the actual temperature, which will also reduce the accuracy of the temperature measurement results of the target power equipment.
- AI technology is a technical discipline that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence. AI technology obtains the best results by sensing the environment, acquiring knowledge and using knowledge.
- artificial intelligence technology is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence.
- Using artificial intelligence for data processing is a common application of artificial intelligence.
- Figure 1 is a structural diagram of the main framework of artificial intelligence.
- the following is an explanation of the above artificial intelligence theme framework from the two dimensions of "intelligent information chain” (horizontal axis) and “IT value chain” (vertical axis).
- the "intelligent information chain” reflects a series of processes from data acquisition to processing. For example, it can be a general process of intelligent information perception, intelligent information representation and formation, intelligent reasoning, intelligent decision-making, intelligent execution and output. In this process, the data has undergone a condensation process of "data-information-knowledge-wisdom".
- the "IT value chain” reflects the value that artificial intelligence brings to the information technology industry from the underlying infrastructure of human intelligence, information (providing and processing technology implementation) to the industrial ecology process of the system.
- the infrastructure provides computing power support for the artificial intelligence system, enables communication with the outside world, and is supported by the basic platform. It communicates with the outside world through sensors; computing power is provided by smart chips (CPU, NPU, GPU, ASIC, FPGA and other hardware acceleration chips); the basic platform includes distributed computing frameworks and networks and other related platform guarantees and support, which can include cloud storage and computing, interconnected networks, etc. For example, sensors communicate with the outside world to obtain data, and these data are provided to the smart chips in the distributed computing system provided by the basic platform for calculation.
- smart chips CPU, NPU, GPU, ASIC, FPGA and other hardware acceleration chips
- the basic platform includes distributed computing frameworks and networks and other related platform guarantees and support, which can include cloud storage and computing, interconnected networks, etc.
- sensors communicate with the outside world to obtain data, and these data are provided to the smart chips in the distributed computing system provided by the basic platform for calculation.
- the data on the upper layer of the infrastructure is used to represent the data sources in the field of artificial intelligence.
- the data involves graphics, images, voice, text, and IoT data of traditional devices, including business data of existing systems and perception data such as force, displacement, liquid level, temperature, and humidity.
- Data processing usually includes data training, machine learning, deep learning, search, reasoning, decision-making and other methods.
- machine learning and deep learning can symbolize and formalize data for intelligent information modeling, extraction, preprocessing, and training. wait.
- Reasoning refers to the process of simulating human intelligent reasoning in computers or intelligent systems, using formalized information to perform machine thinking and solve problems based on reasoning control strategies. Typical functions are search and matching.
- Decision-making refers to the process of making decisions after intelligent information is reasoned, usually providing functions such as classification, sorting, and prediction.
- some general capabilities can be further formed based on the results of the data processing, such as an algorithm or a general system, for example, translation, text analysis, computer vision processing, speech recognition, image recognition, etc.
- Smart products and industry applications refer to the products and applications of artificial intelligence systems in various fields. They are the encapsulation of the overall artificial intelligence solution, which productizes intelligent information decision-making and realizes practical applications. Its application areas mainly include: smart terminals, smart transportation, smart medical care, autonomous driving, smart cities, etc.
- FIG2a is a schematic diagram of a device temperature measurement system provided in an embodiment of the present application, wherein the device temperature measurement system includes a user device and a data processing device.
- the user device includes a smart terminal such as a mobile phone, a personal computer or a robot.
- the user device is the initiator of the device temperature measurement, and as the initiator of the device temperature measurement request, the request is usually initiated by the user through the user device.
- the above-mentioned data processing device can be a device or server with data processing function such as a cloud server, a network server, an application server and a management server.
- the data processing device receives the device temperature measurement request from the smart terminal through the interactive interface, and then performs text processing in the form of machine learning, deep learning, search, reasoning, decision-making, etc. through the memory for storing data and the processor link for data processing.
- the memory in the data processing device can be a general term, including local storage and databases for storing historical data.
- the database can be on the data processing device or on other network servers.
- the user device can receive the user's instruction, and the user device can shoot the target device located in the target area based on the instruction, thereby obtaining an image of the target area, and then initiate a temperature measurement request to the data processing device, so that the data processing device executes an image processing application based on the request for the image obtained by the user device, thereby obtaining an image processing result.
- the user device collects a visible light image of the target area and a thermal infrared image of the target area, and sends a temperature measurement request for these images to the data processing device, so that the data processing device performs a series of analysis and processing on the visible light image of the target area and the thermal infrared image of the target area, thereby obtaining an image processing result, that is, a temperature measurement result of the target device.
- the data processing device may execute the device temperature measurement method of the embodiment of the present application.
- Figure 2b is another structural schematic diagram of the device temperature measurement system provided in an embodiment of the present application.
- the user device directly serves as a data processing device.
- the user device can, under the user's instructions, obtain images from the target area and directly perform image processing applications by the hardware of the user device itself.
- the specific process is similar to that of Figure 2a. Please refer to the above description and will not be repeated here.
- the user device collects a visible light image of the target area and a thermal infrared image of the target area under the user's instruction, and performs a series of analysis and processing on the visible light image of the target area and the thermal infrared image of the target area to obtain the image processing result and the temperature measurement result of the target device.
- the user device itself can execute the device temperature measurement method of the embodiment of the present application.
- FIG. 2c is a schematic diagram of related equipment for device temperature measurement provided in an embodiment of the present application.
- the user device in the above Figures 2a and 2b can specifically be the local device 301 or the local device 302 in Figure 2c
- the data processing device in Figure 2a can specifically be the execution device 210 in Figure 2c
- the data storage system 250 can store the data to be processed of the execution device 210
- the data storage system 250 can be integrated on the execution device 210, and can also be set on the cloud or other network servers.
- the processors in Figures 2a and 2b can perform data training/machine learning/deep learning through a neural network model or other models (for example, a model based on a support vector machine), and use the model finally trained or learned from the data to execute image processing applications on the image, thereby obtaining corresponding processing results.
- a neural network model or other models for example, a model based on a support vector machine
- FIG 3 is a schematic diagram of the system 100 architecture provided in an embodiment of the present application.
- the execution device 110 is configured with an input/output (I/O) interface 112 for data interaction with an external device.
- the user can input data to the I/O interface 112 through the client device 140.
- the input data can include: temperature measurement tasks, task data (including visible light images and thermal infrared images to be processed) and task parameters in the embodiment of the present application.
- the execution device 110 When the execution device 110 preprocesses the input data, or when the computing module 111 of the execution device 110 performs calculation and other related processing (such as implementing the function of the neural network in the present application), the execution device 110 can call the data, code, etc. in the data storage system 150 for the corresponding processing, and can also store the data, instructions, etc. obtained by the corresponding processing in the data storage system 150.
- the I/O interface 112 returns the processing result to the client device 140 so as to provide it to the user.
- the training device 120 can generate corresponding target models/rules based on different training data for different goals or different tasks, and the corresponding target models/rules can be used to achieve the above goals or complete the above tasks, thereby providing the user with the desired results.
- the training data can be stored in the database 130 and come from the training samples collected by the data acquisition device 160.
- the training device 120 and the execution device 110 can be different devices or the same device (that is, the training device 120 is integrated in the execution device 110).
- the user can manually give input data, and the manual giving can be operated through the interface provided by the I/O interface 112.
- the client device 140 can automatically send input data to the I/O interface 112. If the client device 140 is required to automatically send input data and needs to obtain the user's authorization, the user can set the corresponding authority in the client device 140.
- the user can view the results output by the execution device 110 on the client device 140, and the specific presentation form can be a specific method such as display, sound, action, etc.
- the client device 140 can also be used as a data acquisition terminal to collect the input data of the input I/O interface 112 and the output results of the output I/O interface 112 as shown in the figure as new sample data, and store them in the database 130.
- the I/O interface 112 directly stores the input data of the input I/O interface 112 and the output results of the output I/O interface 112 as new sample data in the database 130.
- FIG3 is only a schematic diagram of a system architecture provided by an embodiment of the present application, and the positional relationship between the devices, components, modules, etc. shown in the figure does not constitute any limitation.
- the data storage system 150 is an external memory relative to the execution device 110. In other cases, the data storage system 150 can also be placed in the execution device 110.
- a neural network can be obtained by training according to the training device 120.
- the embodiment of the present application also provides a chip, which includes a neural network processor NPU.
- the chip can be set in the execution device 110 as shown in Figure 3 to complete the calculation work of the calculation module 111.
- the chip can also be set in the training device 120 as shown in Figure 3 to complete the training work of the training device 120 and output the target model/rule.
- Neural network processor NPU is mounted on the main central processing unit (CPU) (host CPU) as a coprocessor, and the main CPU assigns tasks.
- the core part of NPU is the operation circuit, and the controller controls the operation circuit to extract data from the memory (weight memory or input memory) and perform operations.
- the arithmetic circuit includes multiple processing units (process engines, PEs) internally.
- the arithmetic circuit is a two-dimensional systolic array.
- the arithmetic circuit can also be a one-dimensional systolic array or other electronic circuits capable of performing mathematical operations such as multiplication and addition.
- the arithmetic circuit is a general-purpose matrix processor.
- the operation circuit takes the corresponding data of matrix B from the weight memory and caches it on each PE in the operation circuit.
- the operation circuit takes the matrix A data from the input memory and performs matrix operations with matrix B.
- the partial results or final results of the matrix are stored in the accumulator.
- the vector calculation unit can further process the output of the operation circuit, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison, etc.
- the vector calculation unit can be used for network calculations of non-convolutional/non-FC layers in neural networks, such as pooling, batch normalization, local response normalization, etc.
- the vector computation unit can store the processed output vector to a unified buffer.
- the vector computation unit can apply a nonlinear function to the output of the computation circuit, such as a vector of accumulated values, to generate an activation value.
- the vector computation unit generates a normalized value, a merged value, or both.
- the processed output vector can be used as an activation input to the computation circuit, such as for use in a subsequent layer in a neural network.
- the unified memory is used to store input data and output data.
- the weight data is directly transferred from the external memory to the input memory and/or the unified memory through the direct memory access controller (DMAC), the weight data in the external memory is stored in the weight memory, and the data in the unified memory is stored in the external memory.
- DMAC direct memory access controller
- the bus interface unit (BIU) is used to implement interaction between the main CPU, DMAC and instruction fetch memory through the bus.
- An instruction fetch buffer connected to the controller, used to store instructions used by the controller
- the controller is used to call the instructions cached in the memory to control the working process of the computing accelerator.
- the unified memory, input memory, weight memory and instruction fetch memory are all on-chip memories
- the external memory is a memory outside the NPU, which can be a double data rate synchronous dynamic random access memory (DDR SDRAM), a high bandwidth memory (HBM) or other readable and writable memories.
- DDR SDRAM double data rate synchronous dynamic random access memory
- HBM high bandwidth memory
- a neural network may be composed of neural units, and a neural unit may refer to an operation unit with xs and intercept 1 as input, and the output of the operation unit may be:
- n is a natural number greater than 1
- Ws is the weight of xs
- b is the bias of the neural unit.
- f is the activation function of the neural unit, which is used to introduce nonlinear characteristics into the neural network to convert the input signal in the neural unit into the output signal.
- the output signal of the activation function can be used as the input of the next convolutional layer.
- the activation function can be a sigmoid function.
- a neural network is a network formed by connecting many of the above-mentioned single neural units together, that is, the output of one neural unit can be the input of another neural unit.
- the input of each neural unit can be connected to the local receptive field of the previous layer to extract the characteristics of the local receptive field.
- the local receptive field can be an area composed of several neural units.
- the word "space” is used here because the classified object is not a single thing, but a class of things, and space refers to the collection of all individuals of this class of things.
- W is a weight vector, and each value in the vector represents the weight value of a neuron in the neural network of this layer.
- the vector W determines the spatial transformation from input space to output space described above, that is, the weight W of each layer controls how to transform the space.
- the purpose of training a neural network is to finally obtain the weight matrix of all layers of the trained neural network (the weight matrix formed by many layers of vectors W). Therefore, the training process of a neural network is essentially about learning how to control spatial transformations, or more specifically, learning the weight matrix.
- Neural networks can use the error back propagation (BP) algorithm to correct the size of the parameters in the initial neural network model during the training process, so that the reconstruction error loss of the neural network model becomes smaller and smaller. Specifically, the forward transmission of the input signal to the output will generate error loss, and the error loss information is back-propagated to update the parameters in the initial neural network model, so that the error loss converges.
- the back propagation algorithm is a back propagation movement dominated by error loss, which aims to obtain the optimal parameters of the neural network model, such as the weight matrix.
- the model training method provided in the embodiment of the present application involves the processing of data sequences, and can be specifically applied to methods such as data training, machine learning, and deep learning, and symbolizes and formalizes intelligent information modeling, extraction, preprocessing, and training of training data, and finally obtains a trained neural network (such as the first model, the second model, the third model, the fourth model, and the fifth model in the present application, etc.); and the device temperature measurement method provided in the embodiment of the present application can use the above-mentioned trained neural network to input input data (for example, the visible light image of the target area and the thermal infrared image of the target area in the present application, etc.) into the trained neural network to obtain output data (such as the visible light image of the temperature abnormality area of the target device and the thermal infrared image of the temperature abnormality area in the present application, etc.).
- input input data for example, the visible light image of the target area and the thermal infrared image of the target area in the present application, etc.
- output data such as the
- model training method and the device temperature measurement method provided in the embodiment of the present application are inventions based on the same concept, and can also be understood as two parts in a system, or two stages of an overall process: such as the model training stage and the model application stage.
- the equipment temperature measurement method provided in the embodiment of the present application can be applied to the temperature measurement scenario of power equipment.
- the staff of the substation can order the robot to patrol the substation.
- the robot controls the camera to shoot the target power equipment and the target area where the remaining power equipment is located, and uploads the obtained visible light image and thermal infrared image of the target area to the cloud server, so that the cloud server performs a series of processing on these images to obtain the temperature measurement results of the temperature abnormal area in the target power equipment, and feeds back to the staff, so that the staff can inspect the temperature abnormal area of the target power equipment in the substation, thereby maintaining the normal working state of the substation.
- Figure 4 is a flow chart of the equipment temperature measurement method provided in the embodiment of the present application. As shown in Figure 4, the method includes:
- the staff issues instructions to the robot, and the robot can determine the inspection task to be completed based on the instruction, that is, complete the temperature measurement of the target device. Then, in the inspection execution stage (that is, the stage of executing the inspection task), the robot can control the camera to shoot the target area where the target device is located, thereby obtaining a visible light image of the target area and a thermal infrared image of the target area.
- the target area usually refers to an area with a certain range, which contains the target device and the regional equipment.
- the visible light image of the target area contains the original information of each point in the target area (that is, the color, texture, etc. of each point), and the thermal infrared image of the target area contains the temperature information of each point in the target area (that is, the temperature of each point).
- the camera installed on the robot may include an optical imaging camera and a thermal imaging camera, and the optical imaging camera and the thermal imaging camera are installed side by side on the robot's pan-tilt platform, and the pan-tilt platform can achieve 360-degree rotation under the control of the robot. Then, during continuous driving, the robot can control the optical imaging camera to shoot the target area, thereby obtaining a visible light image of the target area, and control the thermal imaging camera to shoot the target area, thereby obtaining a thermal infrared image of the target area.
- the robot can photograph the target area in the following ways to obtain a visible light image of the target area and a thermal infrared image of the target area:
- the camera can be immediately controlled to shoot the target area at a preset angle at this moment, thereby obtaining a visible light image of the target area and a thermal infrared image of the target area.
- the preset position and the preset angle are both determined in advance during the inspection planning stage. Therefore, the position determined in advance should meet the following conditions: when the robot is at this position, the distance between the robot's camera and the target device in the target area is within a preset range (the size of the left and right end values of the range can be set according to actual needs and is not limited here).
- the angle determined in advance meets the following conditions: when the robot's camera shoots the target area at this angle, in the shooting field of view presented, the degree of overlap between the target device in the target area and the remaining devices in the target area is less than the preset overlap threshold (the size of the threshold can be set according to actual needs and is not limited here).
- the inspection planning stage is before the inspection execution stage, and the inspection planning stage is similar to the inspection execution stage. Therefore, the process of the robot completing the inspection planning stage can refer to the process of the robot completing the inspection execution stage (i.e., refer to step 401 and step 402, etc.), which will not be repeated here.
- each time the robot reaches a position it selects an angle and immediately determines at that moment whether the position satisfies the following conditions: when the robot is at that position, whether the distance between the robot's camera and the target device in the target area is within the preset range, and whether a certain angle satisfies the following conditions: when the robot's camera shoots the target area at that angle, whether the overlap between the target device in the target area and the remaining devices in the target area in the presented shooting field of view is less than a preset threshold. If both conditions are met, the robot controls the camera at that position to shoot the target area at that angle, thereby obtaining a visible light image of the target area and a thermal infrared image of the target area. If one of these two conditions is not met, the robot moves to the next position. The above process is repeated for the next position and the next angle until the visible light image and the thermal infrared image of the target area are successfully collected.
- the robot can send the visible light image and the thermal infrared image of the target area to the cloud server, and the cloud server can remove the visible light image and the thermal infrared image of the remaining areas of the target area except the target device from the visible light image and the thermal infrared image of the target area, thereby obtaining the visible light image and the thermal infrared image of the target device.
- the visible light image of the target area presents the entire target area
- the thermal infrared image of the target area presents the entire target area
- the visible light image of the target device only presents the target device
- the thermal infrared image of the target device only presents the target device.
- the cloud server can obtain the visible light image and the thermal infrared image of the target device in the following manner, and the process includes:
- the cloud server can align the visible light image of the target area to the thermal infrared image of the target area to obtain the aligned visible light image of the target area. It is worth noting that the alignment process includes two stages: the initial matching (rough matching) and the secondary matching (fine matching). The following introduces these two stages respectively:
- (1.1) Initial matching stage: After obtaining the visible light image and the thermal infrared image of the target area, the cloud server can crop the visible light image of the target area based on the content matching relationship between the visible light image and the thermal infrared image, so that the size of the cropped visible light image of the target area is roughly similar to the size of the thermal infrared image of the target area (generally, the size of the cropped visible light image of the target area is slightly larger than the size of the thermal infrared image of the target area).
- the content presented by the cropped visible light image of the target area is roughly similar to the content presented by the thermal infrared image of the target area (generally, the content presented by the cropped visible light image of the target area is slightly more than the content presented by the thermal infrared image of the target area).
- the staff in the inspection planning stage, can hold the calibration plate, which includes a calibration plate support panel, a heating panel and a hollow circular dot matrix glass panel stacked in sequence, wherein the hollow circular dot matrix glass panel is located on the top layer, the calibration plate support panel is located on the bottom layer, and the heating panel located in the middle layer can be heated by a temperature controller.
- the staff can control the robot to shoot the calibration plate through the camera, thereby obtaining the visible light image of the calibration plate and the thermal infrared image of the calibration plate, and upload them to the cloud server.
- Figures 6a, 6b and 6c Figure 6a is a schematic diagram of the calibration process provided in an embodiment of the present application
- Figure 6b is another schematic diagram of the calibration process provided in an embodiment of the present application
- Figure 6c is another schematic diagram of the calibration process provided in an embodiment of the present application
- the cloud server can determine the positions of the multiple hollow circular dots in the visible light image of the calibration plate and the positions in the thermal infrared image of the calibration plate, and based on the positions of the multiple hollow circular dots in these two images, it can be determined which part of the visible light image of the calibration plate cannot find a matching image in the thermal infrared image of the calibration plate (that is, the content presented by this part of the image in the visible light image does not appear
- the cloud server can input the cropped visible light image of the target area and the thermal infrared image of the target area into the first model.
- Figure 7 is a schematic diagram of the refined matching provided in an embodiment of the present application
- the first model can calculate the distance between a certain pixel in the cropped visible light image of the target area and all the pixels in the thermal infrared image of the target area, and use the pixel with the smallest distance in the thermal infrared image as the pixel that matches the pixel in the visible light image.
- the first model can also perform the same operation as that performed on the pixel, so the first model can obtain the pixel matching relationship between the cropped visible light image of the target area and the thermal infrared image of the target area. Based on the pixel matching relationship, the first model can match all the pixels in the cropped visible light image of the target area.
- the cropped visible light image of the target area can be accurately projected onto the thermal infrared image of the target area, and the projected visible light image is the aligned visible light image of the target area.
- the first model is a trained neural network model, and its structure is shown in Figure 8 ( Figure 8 is a structural diagram of the first model provided in an embodiment of the present application).
- the first model includes: a VGG network module, a position information encoding module, an attention module, an upsampling module, a feature processing module, and a projection transformation module.
- the cloud server may first obtain the first model to be trained (i.e., the neural network model to be trained) and a batch of training data (including a cropped visible light image of a certain area and a thermal infrared image of the area), and the actual processing result of the training data (i.e., the aligned actual visible light image of the area). Then, the cloud server may input the training data into the first model to be trained so as to process the training data through the first model to be trained and obtain the estimated processing result of the training data (i.e., the aligned estimated visible light image of the area).
- the first model to be trained i.e., the neural network model to be trained
- a batch of training data including a cropped visible light image of a certain area and a thermal infrared image of the area
- the actual processing result of the training data i.e., the aligned actual visible light image of the area.
- the cloud server may input the training data into the first model to be trained so as to process the training data through
- the cloud server may calculate the actual processing result of the training data and the estimated processing result of the training data through a preset first loss function to obtain the target loss, which is used to indicate the difference between the actual processing result of the training data and the estimated processing result of the training data.
- the cloud server may update the parameters of the first model to be trained based on the target loss, and continue to train the first model to be trained after the updated parameters using the next batch of training data until the model training conditions are met (e.g., the target loss converges, etc.), thereby obtaining the first model.
- the cloud server After obtaining the aligned visible light image of the target area, the cloud server also calculates the thermal infrared image of the target area and the aligned visible light image to obtain a depth image of the target area.
- the depth image of the target area contains the depth information of each point in the target area (i.e., the depth of each point, etc.).
- the calculation process includes:
- the cloud server can convert the aligned visible light image of the target area into a pseudo infrared image of the target area, and input the pseudo infrared image of the target area and the thermal infrared image of the target area into the second model, so that the second model can perform a series of calculations on these images to obtain a disparity map, wherein the disparity map includes the disparity between each pixel in the pseudo infrared image and the corresponding pixel in the thermal infrared image.
- the second model is a trained neural network model, and its structure is shown in Figure 9 ( Figure 9 is a structural schematic diagram of the second model provided in an embodiment of the present application).
- the second model may include: a pseudo-infrared image individual feature extraction module, a pseudo-infrared image and thermal infrared image common feature extraction module, a thermal infrared image individual feature extraction module, a feature fusion module and a high-resolution reconstruction module.
- the cloud server may first obtain the second model to be trained (i.e., the neural network model to be trained) and a batch of training data (including a pseudo-infrared image of a certain area and a thermal infrared image of the area), and the actual processing result of the training data (i.e., the actual disparity map). Then, the cloud server may input the training data into the second model to be trained so as to process the training data through the second model to be trained and obtain the estimated processing result of the training data (i.e., the estimated disparity map).
- the second model to be trained i.e., the neural network model to be trained
- a batch of training data including a pseudo-infrared image of a certain area and a thermal infrared image of the area
- the actual processing result of the training data i.e., the actual disparity map
- the cloud server may calculate the actual processing result of the training data and the estimated processing result of the training data through a preset second loss function to obtain the target loss, which is used to indicate the difference between the actual processing result of the training data and the estimated processing result of the training data.
- the cloud server may update the parameters of the second model to be trained based on the target loss, and continue to train the second model to be trained after the updated parameters using the next batch of training data until the model training conditions are met (e.g., the target loss converges, etc.), thereby obtaining the second model.
- the cloud server can calculate the coordinates of the pixel in the pseudo infrared image and the disparity between the pixel in the pseudo infrared image and the corresponding pixel in the thermal infrared image, thereby obtaining the coordinates of the corresponding pixel in the thermal infrared image. Then, the cloud server can calculate the internal and external parameters of the camera, the coordinates of the pixel, and the coordinates of the corresponding pixel to determine the depth corresponding to the pixel.
- the cloud server can also perform the same operation as that performed on the pixel, thereby obtaining the depth corresponding to all pixels of the pseudo infrared image, that is, the depth of all points in the target area.
- the depth of all points in the target area constitutes the depth image of the target area.
- the cloud server can obtain the optical center coordinates Oir of the optical imaging camera, the optical center coordinates Orgb of the thermal imaging camera, the difference between the two optical center coordinates is T, the focal length fir of the optical imaging camera, and the focal length frgb of the thermal imaging camera.
- the depth Z corresponding to the pixel point can be calculated by the following formula:
- the depths corresponding to all pixels can be calculated to obtain a depth image of the target area.
- the cloud server can determine the thermal infrared image of the foreground in the target area and the visible light image of the foreground in the target area based on the depth image of the target area, from the thermal infrared image of the target area and the aligned visible light image of the target area.
- the visible light image of the foreground in the target area only presents the foreground of the target area
- the thermal infrared image of the foreground in the target area only presents the foreground of the target area.
- the foreground of the target area includes the target device and other devices, and the background of the target area is the environment where the target device is located (for example, the sky in the distance, a mountain peak, etc.).
- the determination process includes:
- the cloud server can determine the depth corresponding to all pixels in the thermal infrared image of the target area and the depth corresponding to all pixels in the aligned visible light image of the target area based on the depth image of the target area.
- the cloud server can remove pixels whose depth is greater than a preset depth threshold (the size of the threshold can be set according to actual needs and is not limited here), which is equivalent to removing the thermal infrared image of the background in the target area and retaining the thermal infrared image of the foreground in the target area.
- the cloud server can remove pixels whose depth is greater than a preset depth threshold, which is equivalent to removing the visible light image of the background in the target area and retaining the visible light image of the foreground in the target area.
- the cloud server can segment the thermal infrared image of the foreground of the target area and the visible light image of the foreground of the target area based on the depth image of the target area to obtain the visible light image of the target sub-area and the thermal infrared image of the target sub-area.
- the target sub-area also called the target detection frame
- the segmentation process includes:
- the cloud server may input the depth image of the target area, the thermal infrared image of the foreground in the target area, and the visible light image of the foreground in the target area into the third model, so that the third model, based on the depth image of the target area, removes the thermal infrared image of the remaining areas in the foreground except the target sub-area from the thermal infrared image of the foreground in the target area, and retains the thermal infrared image of the target sub-area.
- the third model may also remove the visible light image of the remaining areas in the foreground except the target sub-area from the visible light image of the foreground in the target area, and retain the visible light image of the target sub-area.
- the third model is a trained neural network model, and its structure is shown in Figure 11 ( Figure 11 is a structural schematic diagram of the third model provided in an embodiment of the present application).
- the third model may include: a visible light image feature encoder, a thermal infrared image feature encoder, a feature fusion module, a feature fusion module based on a dual-space graph, a decoder, a convolution kernel (1*1 convolution kernel and 3*3 convolution kernel, etc.) and an upsampling module.
- the cloud server may first obtain the third model to be trained (i.e., the neural network model to be trained) and a batch of training data (including a depth image of a certain area, a thermal infrared image of the foreground of the area, and a thermal infrared image of the foreground of the area).
- the cloud server may input the training data into the third model to be trained, so as to process the training data through the third model to be trained, and obtain the estimated processing results of the training data (i.e., the estimated visible light image of the detection box and the estimated thermal infrared image).
- the cloud server may calculate the actual processing results of the training data and the estimated processing results of the training data through a preset third loss function to obtain the target loss, and the target loss is used to indicate the difference between the actual processing results of the training data and the estimated processing results of the training data.
- the cloud server may update the parameters of the third model to be trained based on the target loss, and continue to train the third model to be trained with the updated parameters using the next batch of training data until the model training conditions are met (for example, the target loss converges, etc.), thereby obtaining the third model.
- the cloud server may perform secondary segmentation on the visible light image and the thermal infrared image of the target sub-region, thereby obtaining the visible light image and the thermal infrared image of the target device.
- the secondary segmentation process includes:
- the cloud server can determine the temperature corresponding to each pixel in the visible light image of the target sub-region based on the thermal infrared image of the target sub-region. Since the cloud server has determined the device type to which the target device belongs, the temperature range of the target device can be obtained. Then, the cloud server can remove the pixels whose temperature is outside the temperature range in the visible light image of the target sub-region, which is equivalent to removing the visible light images of a part of some other devices in the target sub-region, retaining a part of other other devices and the visible light image of the target device.
- the retained part of the visible light image can be called the visible light image of the optimized sub-region (i.e., the optimized detection frame).
- the cloud server can remove the pixels whose temperature is outside the temperature range in the thermal infrared image of the target sub-region, which is equivalent to removing the thermal infrared images of a part of some other devices in the target sub-region, retaining a part of other other devices and the thermal infrared image of the target device.
- the retained part of the thermal infrared image can be called the thermal infrared image of the optimized sub-region.
- the cloud server can input the visible light image of the optimized sub-region and the thermal infrared image of the optimized sub-region into the fourth model, so that the fourth model removes the visible light images of some other devices from the visible light image of the optimized sub-region and retains the visible light image of the target device.
- the fourth model can also remove the thermal infrared images of some other devices from the thermal infrared image of the optimized sub-region and retain the thermal infrared image of the target device.
- the fourth model is a trained neural network model, and its structure is shown in Figure 12 ( Figure 12 is a structural diagram of the fourth model provided in an embodiment of the present application).
- the fourth model may include: a visible light image encoder, a thermal infrared image encoder, a global attention module, a visible light image decoder, a thermal infrared image decoder, and a convolution module (1-channel convolution).
- the cloud server may first obtain the fourth model to be trained (i.e., the neural network model to be trained) and a batch of training data (including a thermal infrared image of a detection frame and a visible light image of the detection frame), and the actual processing results of the training data (i.e., the actual visible light image and the actual thermal infrared image of a device). Then, the cloud server may input the training data into the fourth model to be trained, so as to process the training data through the fourth model to be trained and obtain the estimated processing results of the training data (i.e., the estimated visible light image and the estimated thermal infrared image of the device).
- the fourth model to be trained i.e., the neural network model to be trained
- a batch of training data including a thermal infrared image of a detection frame and a visible light image of the detection frame
- the actual processing results of the training data i.e., the actual visible light image and the actual thermal infrared image of a device.
- the cloud server may calculate the actual processing results of the training data and the estimated processing results of the training data through the preset fourth loss function to obtain the target loss, which is used to indicate the difference between the actual processing results of the training data and the estimated processing results of the training data.
- the cloud server may update the parameters of the fourth model to be trained based on the target loss, and continue to train the fourth model to be trained after the updated parameters using the next batch of training data until the model training conditions are met (e.g., the target loss converges, etc.), thereby obtaining the fourth model.
- the cloud server may execute all the steps or selectively execute some of the steps.
- the cloud server may only execute steps (1), (2), (4) and (5), that is, the cloud server first aligns the visible light image of the target area to the thermal infrared image of the target area to obtain the aligned visible light image of the target area.
- the cloud server calculates the thermal infrared image of the target area and the aligned visible light image to obtain the depth image of the target area.
- the cloud server may segment the visible light image of the target area and the thermal infrared image of the target area based on the depth image of the target area to obtain the visible light image of the target sub-area and the thermal infrared image of the target sub-area, where the target sub-area is the area occupied by the target device and part of the remaining devices in the target area. Finally, the cloud server performs secondary segmentation on the visible light image of the target sub-area and the thermal infrared image of the target sub-area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server can determine the visible light image and the thermal infrared image of the temperature abnormality region in the target device from the visible light image and the thermal infrared image of the target device. It is understandable that the visible light image of the temperature abnormality region in the target device only presents the temperature abnormality region, and the thermal infrared image of the temperature abnormality region only presents the temperature abnormality region.
- the elimination process includes: after obtaining the visible light image and the thermal infrared image of the target device, the cloud server can input the visible light image and the thermal infrared image of the target device into the fifth model, so that the fifth model eliminates the visible light image of the normal temperature area of the target device in the visible light image of the target device, and retains the visible light image of the abnormal temperature area of the target device.
- the fifth model can also eliminate the thermal infrared image of the normal temperature area in the thermal infrared image of the target device, and retain the thermal infrared image of the abnormal temperature area.
- the fifth model is a trained neural network model, and its structure is shown in Figure 13 ( Figure 13 is a structural diagram of the fifth model provided in an embodiment of the present application).
- the fifth model may include: a preprocessing module, an encoder module, a multimodal feature aggregation module, a global attention module, a decoder module, a convolution module and an upsampling module.
- the cloud server may first obtain the fifth model to be trained (i.e., the neural network model to be trained) and a batch of training data (including a thermal infrared image of a certain device and a visible light image of the device), and the actual processing results of the training data (i.e., the actual visible light image and the actual thermal infrared image of the temperature abnormality area in the device). Then, the cloud server may input the training data into the fifth model to be trained, so as to process the training data through the fifth model to be trained, and obtain the estimated processing results of the training data (i.e., the estimated visible light image and the estimated thermal infrared image of the temperature abnormality area in the device).
- the fifth model to be trained i.e., the neural network model to be trained
- a batch of training data including a thermal infrared image of a certain device and a visible light image of the device
- the actual processing results of the training data i.e., the actual visible light image and the actual thermal infrared image of
- the cloud server may calculate the actual processing results of the training data and the estimated processing results of the training data through the preset fifth loss function to obtain the target loss, and the target loss is used to indicate the difference between the actual processing results of the training data and the estimated processing results of the training data.
- the cloud server may update the parameters of the fifth model to be trained based on the target loss, and continue to train the fifth model to be trained after the updated parameters using the next batch of training data until the model training conditions are met (e.g., the target loss converges, etc.), thereby obtaining the fifth model.
- the cloud server can directly use the visible light image and the thermal infrared image of the temperature abnormality area as the temperature measurement result of the temperature abnormality area of the target device and feed it back to the staff, so that the staff can find the temperature abnormality area of the target device based on the temperature measurement result and inspect the temperature abnormality area of the target device.
- the cloud server is also provided with a preset corresponding relationship, which is used to indicate the corresponding relationship between the distance and the temperature correction value.
- the cloud server can first obtain the distance between the thermal imaging camera and the target device (including the distance between the thermal imaging camera and the temperature abnormal area in the target device) from the depth image of the target area, and then determine the temperature correction value corresponding to the distance between the thermal imaging camera and the target device based on the preset corresponding relationship and the distance between the thermal imaging camera and the target device, and then adjust the thermal infrared image of the temperature abnormal area based on the temperature correction value (for example, superimpose the temperature correction value on the thermal infrared image of the temperature abnormal area) to obtain the adjusted thermal infrared image of the temperature abnormal area.
- the cloud server can use the visible light image of the temperature abnormal area and the adjusted thermal infrared image of the temperature abnormal area as the adjusted temperature measurement result of the temperature abnormal area of the target device, and feedback it to the staff.
- the robot under the instruction of the staff, can control the camera to shoot the target area including the target device and the remaining devices, obtain the visible light image of the target area and the thermal infrared image of the target area, and send them to the cloud server.
- the cloud server can process the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server can continue to process the visible light image of the target device and the thermal infrared image of the target device to obtain the thermal infrared image of the temperature abnormal area in the target device.
- the cloud server can determine the temperature measurement result of the temperature abnormal area of the target device based on the thermal infrared image of the temperature abnormal area, and report it to the staff so that the staff can repair the temperature abnormal area of the target device.
- the cloud server can remove the visible light image and thermal infrared image of the remaining devices from them to obtain the visible light image and thermal infrared image of the target device. Since the visible light image and thermal infrared image of the target device only show the target device, the cloud server will not be affected by other devices during the analysis of these images.
- the visible light image and thermal infrared image of the temperature abnormal area in the target device can be accurately identified as the temperature abnormal area of the target device. It can be seen that the cloud server accurately determines the temperature abnormal area of the target device without misjudgment, thereby effectively improving the accuracy of the temperature measurement results of the target device.
- the cloud server can select a preset position and preset angle for the robot based on the images of the target area taken by the robot at different positions and angles according to certain conditions. These conditions include: when the robot is at this position, the distance between the camera and the target device in the target area is within a preset range, and when the robot controls the camera to shoot the target area at this angle, the overlap between the target device and the remaining devices is less than a preset threshold. It can be seen that the cloud server can automatically plan the optimal position and angle for the robot according to the aforementioned conditions, and the factors considered are relatively comprehensive. In this way, the robot can take the optimal image at the optimal position and angle during the inspection execution stage, which is not only efficient, but also improves the accuracy of the temperature measurement results of the target device.
- the cloud server can determine the temperature correction value corresponding to the distance between the target device and the camera, so as to adjust the thermal infrared image of the temperature abnormality area and obtain the adjusted thermal infrared image of the temperature abnormality area.
- the adjusted thermal infrared image of the temperature abnormality area in the target device is closer to the actual temperature of the temperature abnormality area in the target device, which is conducive to improving the accuracy of the temperature measurement result of the target device.
- the cloud server can perform a series of calculations on the visible light image and thermal infrared image of the target area to obtain the depth image of the target area. Since the depth image of the target area contains the distance between each point in the target area and the camera, the cloud server can automatically obtain the distance between the camera and the temperature abnormality area in the target device based on the depth image of the target area, so as to determine the temperature correction value corresponding to the distance, and use the temperature correction value to complete the adjustment of the temperature measurement result. It can be seen that the entire process of temperature correction can be automatically completed by the cloud server, without the need for staff to operate, which can reduce the cost of manual operation.
- FIG. 14 is a structural schematic diagram of the device temperature measurement device provided in the embodiment of the present application. As shown in FIG. 14 , the device includes:
- the shooting module 1401 is used to control the camera to shoot the target area to obtain a visible light image and a thermal infrared image of the target area, where the target area includes the target device;
- a first processing module 1402 is used to process the visible light image and the thermal infrared image of the target area to obtain a visible light image and a thermal infrared image of the target device;
- the second processing module 1403 is used to process the visible light image of the target device and the thermal infrared image of the target device to obtain a thermal infrared image of the temperature abnormality area in the target device;
- the determination module 1404 is used to determine the temperature measurement result of the temperature abnormality area based on the thermal infrared image of the temperature abnormality area.
- the robot under the instruction of the staff, can control the camera to shoot the target area including the target device and the remaining devices, obtain the visible light image of the target area and the thermal infrared image of the target area, and send them to the cloud server.
- the cloud server can process the visible light image of the target area and the thermal infrared image of the target area to obtain the visible light image of the target device and the thermal infrared image of the target device.
- the cloud server can continue to process the visible light image of the target device and the thermal infrared image of the target device to obtain the thermal infrared image of the temperature abnormal area in the target device.
- the cloud server can determine the temperature measurement result of the temperature abnormal area of the target device based on the thermal infrared image of the temperature abnormal area, and report it to the staff so that the staff can repair the temperature abnormal area of the target device.
- the cloud server can remove the visible light image and thermal infrared image of the remaining devices from them to obtain the visible light image and thermal infrared image of the target device.
- the cloud server Since the visible light image and thermal infrared image of the target device only present the target device, the cloud server will not be affected by other devices during the analysis of these images, and accurately confirm the visible light image and thermal infrared image of the temperature abnormal area in the target device as the temperature measurement result of the temperature abnormal area of the target device. It can be seen that the cloud server accurately determines the temperature abnormal area of the target device without misjudgment, thereby effectively improving the accuracy of the temperature measurement results of the target device.
- the device also includes: a computing module, used to calculate the thermal infrared image of the target area and the visible light image of the target area to obtain a depth image of the target area; a first processing module 1402, used to process the depth image of the target area, the visible light image of the target area and the thermal infrared image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- a computing module used to calculate the thermal infrared image of the target area and the visible light image of the target area to obtain a depth image of the target area
- a first processing module 1402 used to process the depth image of the target area, the visible light image of the target area and the thermal infrared image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- the device further includes: a third processing module, configured to align the visible light image of the target area with the thermal infrared image of the target area to obtain the aligned visible light image of the target area; and a computing module, configured to compute the thermal infrared image of the target area. and calculating the aligned visible light image to obtain a depth image of the target area; a first processing module 1402 is used to process the aligned visible light image of the target area, the thermal infrared image of the target area and the depth image of the target area to obtain a visible light image of the target device and a thermal infrared image of the target device.
- a third processing module configured to align the visible light image of the target area with the thermal infrared image of the target area to obtain the aligned visible light image of the target area
- a computing module configured to compute the thermal infrared image of the target area. and calculating the aligned visible light image to obtain a depth image of the target area
- a first processing module 1402 is used to process
- the device also includes: a fourth processing module, which is used to determine, based on the depth image of the target area, the thermal infrared image of the target area and the aligned visible light image, a thermal infrared image of the foreground in the target area and a visible light image of the foreground, wherein the foreground includes a target device; and a first processing module, which is used to process the depth image of the target area, the thermal infrared image of the foreground and the visible light image of the foreground to obtain a visible light image of the target device and a thermal infrared image of the target device.
- a fourth processing module which is used to determine, based on the depth image of the target area, the thermal infrared image of the target area and the aligned visible light image, a thermal infrared image of the foreground in the target area and a visible light image of the foreground, wherein the foreground includes a target device
- a first processing module which is used to process the depth
- the first processing module 1402 is used to: segment the visible light image and the thermal infrared image of the target area to obtain a visible light image and a thermal infrared image of the target sub-area, where the target sub-area is the area occupied by the target device and a part of the remaining devices in the target area; perform secondary segmentation on the visible light image and the thermal infrared image of the target sub-area to obtain a visible light image and a thermal infrared image of the target device.
- the device also includes: an acquisition module, used to acquire the distance between the camera and the temperature abnormality area based on the visible light image of the target area and the thermal infrared image of the target area; an adjustment module, used to adjust the temperature measurement result based on the distance between the camera and the temperature abnormality area and a preset correspondence relationship to obtain an adjusted temperature measurement result of the temperature abnormality area, and the preset correspondence relationship is used to indicate the correspondence between the distance and the temperature correction value.
- an acquisition module used to acquire the distance between the camera and the temperature abnormality area based on the visible light image of the target area and the thermal infrared image of the target area
- an adjustment module used to adjust the temperature measurement result based on the distance between the camera and the temperature abnormality area and a preset correspondence relationship to obtain an adjusted temperature measurement result of the temperature abnormality area, and the preset correspondence relationship is used to indicate the correspondence between the distance and the temperature correction value.
- the shooting module 1401 is used to control the camera to shoot the target area at a preset position and at a preset angle to obtain a visible light image of the target area and a thermal infrared image of the target area.
- the distance between the camera at the position and the target device in the target area is within a preset range.
- the degree of overlap between the target device and other devices is less than a preset threshold.
- the camera includes an optical imaging camera and a thermal imaging camera.
- FIG. 15 is a structural schematic diagram of the execution device provided by the embodiment of the present application.
- the execution device 1500 can be specifically manifested as a mobile phone, a tablet, a laptop computer, an intelligent wearable device, a server, etc., which is not limited here.
- the execution device 1500 can be deployed with the device temperature measuring device described in the embodiment corresponding to FIG. 14, which is used to realize the function of device temperature measurement in the embodiment corresponding to FIG. 4.
- the execution device 1500 includes: a receiver 1501, a transmitter 1502, a processor 1503 and a memory 1504 (wherein the number of processors 1503 in the execution device 1500 can be one or more, and FIG.
- the processor 1503 may include an application processor 15031 and a communication processor 15032.
- the receiver 1501, the transmitter 1502, the processor 1503 and the memory 1504 may be connected via a bus or other means.
- the memory 1504 may include a read-only memory and a random access memory, and provides instructions and data to the processor 1503. A portion of the memory 1504 may also include a non-volatile random access memory (NVRAM).
- NVRAM non-volatile random access memory
- the memory 1504 stores processor and operation instructions, executable modules or data structures, or subsets thereof, or extended sets thereof, wherein the operation instructions may include various operation instructions for implementing various operations.
- the processor 1503 controls the operation of the execution device.
- the various components of the execution device are coupled together through a bus system, wherein the bus system includes not only a data bus but also a power bus, a control bus, and a status signal bus, etc.
- the bus system includes not only a data bus but also a power bus, a control bus, and a status signal bus, etc.
- various buses are referred to as bus systems in the figure.
- the method disclosed in the above embodiment of the present application can be applied to the processor 1503, or implemented by the processor 1503.
- the processor 1503 can be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 1503.
- the above processor 1503 can be a general-purpose processor, a digital signal processor (digital signal processing, DSP), a microprocessor or a microcontroller, and can further include an application specific integrated circuit (application specific integrated circuit, ASIC), a field programmable gate array (field-programmable gate array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
- the processor 1503 can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application.
- the general processor can be a microprocessor or the processor can also be any conventional processor, etc.
- the steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor.
- the software module can be located in a random access memory.
- the processor 1503 reads the information in the memory 1504 and completes the steps of the above method in combination with its hardware.
- the receiver 1501 can be used to receive input digital or character information and generate signal input related to the relevant settings and function control of the execution device.
- the transmitter 1502 can be used to output digital or character information through the first interface; the transmitter 1502 can also be used to send instructions to the disk group through the first interface to modify the data in the disk group; the transmitter 1502 can also include a display device such as a display screen.
- the processor 1503 is used to complete the temperature measurement operation for the target device through the various neural network models (including the first model, the second model, the third model, the fourth model and the fifth model, etc.) in the embodiment corresponding to Figure 4.
- FIG. 16 is a structural schematic diagram of the training device provided by the embodiment of the present application.
- the training device 1600 is implemented by one or more servers.
- the training device 1600 may have relatively large differences due to different configurations or performances, and may include one or more central processing units (CPU) 1614 (for example, one or more processors) and a memory 1632, and one or more storage media 1630 (for example, one or more mass storage devices) storing application programs 1642 or data 1644.
- the memory 1632 and the storage medium 1630 can be short-term storage or permanent storage.
- the program stored in the storage medium 1630 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the training device. Furthermore, the central processor 1614 can be configured to communicate with the storage medium 1630 to execute a series of instruction operations in the storage medium 1630 on the training device 1600.
- the training device 1600 may also include one or more power supplies 1626, one or more wired or wireless network interfaces 1650, one or more input and output interfaces 1658; or, one or more operating systems 1641, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM, etc.
- the training device can implement the training process of each neural network model (including the first model, the second model, the third model, the fourth model and the fifth model, etc.) in the embodiment corresponding to Figure 4.
- each neural network model including the first model, the second model, the third model, the fourth model and the fifth model, etc.
- An embodiment of the present application also relates to a computer storage medium, in which a program for signal processing is stored.
- the program When the program is run on a computer, the computer executes the steps executed by the aforementioned execution device, or the computer executes the steps executed by the aforementioned training device.
- An embodiment of the present application also relates to a computer program product, which stores instructions, which, when executed by a computer, enable the computer to execute the steps executed by the aforementioned execution device, or enable the computer to execute the steps executed by the aforementioned training device.
- the execution device, training device or terminal device provided in the embodiments of the present application may specifically be a chip, and the chip includes: a processing unit and a communication unit, wherein the processing unit may be, for example, a processor, and the communication unit may be, for example, an input/output interface, a pin or a circuit, etc.
- the processing unit may execute the computer execution instructions stored in the storage unit so that the chip in the execution device executes the data processing method described in the above embodiment, or so that the chip in the training device executes the data processing method described in the above embodiment.
- the storage unit is a storage unit in the chip, such as a register, a cache, etc.
- the storage unit may also be a storage unit located outside the chip in the wireless access device end, such as a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM), etc.
- ROM read-only memory
- RAM random access memory
- FIG. 17 is a schematic diagram of the structure of a chip provided in an embodiment of the present application.
- the chip can be expressed as a neural network processor NPU 1700.
- NPU 1700 is mounted on the host CPU (Host CPU) as a coprocessor, and tasks are assigned by the Host CPU.
- the core part of the NPU is the operation circuit 1703, which is controlled by the controller 1704 to extract matrix data from the memory and perform multiplication operations.
- the operation circuit 1703 includes multiple processing units (Process Engine, PE) inside.
- the operation circuit 1703 is a two-dimensional systolic array.
- the operation circuit 1703 can also be a one-dimensional systolic array or other electronic circuits capable of performing mathematical operations such as multiplication and addition.
- the operation circuit 1703 is a general-purpose matrix processor.
- the operation circuit takes the corresponding data of matrix B from the weight memory 1702 and caches it on each PE in the operation circuit.
- the operation circuit takes the matrix A data from the input memory 1701 and performs matrix operation with matrix B, and the partial result or final result of the matrix is stored in the accumulator 1708.
- the unified memory 1706 is used to store input data and output data.
- the weight data is directly transferred to the weight memory 1702 through the direct memory access controller (DMAC) 1705.
- the input data is also transferred to the unified memory 1706 through the DMAC.
- DMAC direct memory access controller
- BIU stands for Bus Interface Unit, that is, bus interface unit 1713, which is used for the interaction between AXI bus and DMAC and instruction fetch buffer (IFB) 1709.
- IOB instruction fetch buffer
- the bus interface unit 1713 (Bus Interface Unit, BIU for short) is used for the instruction fetch memory 1709 to obtain instructions from the external memory, and is also used for the storage unit access controller 1705 to obtain the original data of the input matrix A or the weight matrix B from the external memory.
- BIU Bus Interface Unit
- DMAC is mainly used to transfer input data in the external memory DDR to the unified memory 1706 or to transfer weight data to the weight memory 1702 or to transfer input data to the input memory 1701.
- the vector calculation unit 1707 includes multiple operation processing units, which further process the output of the operation circuit 1703 when necessary, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison, etc. It is mainly used for non-convolutional/fully connected layer network calculations in neural networks, such as Batch Normalization, pixel-level summation, upsampling of predicted label planes, etc.
- the vector calculation unit 1707 can store the processed output vector to the unified memory 1706.
- the vector calculation unit 1707 can apply a linear function; or, a nonlinear function to the output of the operation circuit 1703, such as linear interpolation of the predicted label plane extracted by the convolution layer, and then, for example, a vector of accumulated values to generate an activation value.
- the vector calculation unit 1707 generates a normalized value, a pixel-level summed value, or both.
- the processed output vector can be used as an activation input to the operation circuit 1703, for example, for use in a subsequent layer in a neural network.
- An instruction fetch buffer 1709 connected to the controller 1704, for storing instructions used by the controller 1704;
- Unified memory 1706, input memory 1701, weight memory 1702 and instruction fetch memory 1709 are all on-chip memories. External memories are private to the NPU hardware architecture.
- the processor mentioned in any of the above places may be a general-purpose central processing unit, a microprocessor, an ASIC, or one or more integrated circuits for controlling the execution of the above program.
- the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.
- the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.
- the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.
- a computer device which can be a personal computer, a training device, or a network device, etc.
- all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof.
- all or part of the embodiments may be implemented in the form of a computer program product.
- the computer program product includes one or more computer instructions.
- the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
- the computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.
- the available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.
- a magnetic medium e.g., a floppy disk, a hard disk, a tape
- an optical medium e.g., a DVD
- a semiconductor medium e.g., a solid-state drive (SSD)
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Radiation Pyrometers (AREA)
- Image Processing (AREA)
Abstract
一种设备测温方法及其相关设备,可准确判定目标设备的温度异常区域,从而有效提高目标设备的温度测量结果的准确度。设备测温方法包括:在工作人员的指令下,机器人可控制摄像头对包含目标设备和其余设备的目标区域进行拍摄,得到目标区域的可见光图像和热红外图像,并发送至云服务器。接着,云服务器可对目标区域的可见光图像和热红外图像进行处理,从而得到目标设备的可见光图像和热红外图像。然后,云服务器可继续对目标设备的可见光图像和热红外图像进行处理,得到目标设备中温度异常区域的热红外图像。最后,云服务器可基于温度异常区域的热红外图像,确定目标设备的温度异常区域的温度测量结果,并上报给工作人员。
Description
本申请要求于2022年11月15日提交国家知识产权局、申请号为202211427331.4、发明名称为“一种设备测温方法及其相关设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本申请实施例涉及设备检测技术领域,尤其涉及一种设备测温方法及其相关设备。
变电站中电力设备的预防性检测是维护电力设备的重要手段,测量电力设备的温度又是预防性检测中的重要环节,通过测量电力设备的温度可以确定电力设备中的温度正常区域以及温度异常区域,从而判断电力设备是否能够正常工作,即电力设备是否存在故障。
目前,变电站中的机器人在巡检时,可拍摄包含目标电力设备的某一个区域的热红外图像,并将该图像发送至远端的云服务器进行图像分析。由于该图像包含该区域中各个点的温度,故云服务器基于该图像可确定目标电力设备的温度正常区域以及温度异常区域,以此作为目标电力设备的温度测量结果并及时通知变电站的工作人员,以使得工作人员对目标电力设备的温度异常区域进行检修。
然而,变电站中的电力设备很密集,该区域有可能不仅包含目标电力设备,还包含其余电力设备,云服务器在对该区域的热红外图像进行分析时,容易受到其余电力设备的影响,将其余电力设备边缘附近的某些点的温度误认为目标电力设备中某些点的温度,误判了目标电力设备的温度异常区域,导致针对目标电力设备的温度测量结果的准确度较低。
发明内容
本申请实施例提供了一种设备测温方法及其相关设备,可准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
本申请实施例的第一方面提供了一种设备测温方法,该方法包括:
工作人员对机器人下发指令,机器人基于该指令可确定需要完成的巡检任务,即完成针对目标设备的温度测量。那么,在执行巡检任务的阶段中,机器人可控制摄像头对目标设备所在的目标区域进行拍摄,从而得到目标区域的可见光图像以及目标区域的热红外图像,可以理解的是,目标区域的可见光图像呈现了整个目标区域,目标区域的热红外图像呈现了整个目标区域,目标区域通常指被圈定某个范围的区域,该区域内包含目标设备以及区域设备。
得到目标区域的可见光图像以及目标区域的热红外图像后,机器人可将目标区域的可见光图像以及目标区域的热红外图像发送至云服务器,云服务器可在目标区域的可见光图像和目标区域的热红外图像中,将目标区域中除目标设备之外的其余区域的可见光图像以及其余区域的热红外图像剔除,从而得到目标设备的可见光图像和目标设备的热红外图像。可以理解的是,目标设备的可见光图像仅呈现目标设备,目标设备的热红外图像仅呈现目标设备。
在得到目标设备的可见光图像和目标设备的热红外图像后,云服务器可在目标设备的可见光图像和目标设备的热红外图像中,确定目标设备中温度异常区域的可见光图像和温度异常区域的热红外图像。可以理解的是,目标设备中温度异常区域的可见光图像仅呈现温度异常区域,温度异常区域的热红外图像仅呈现温度异常区域。
在得到目标设备中温度异常区域的可见光图像和温度异常区域的热红外图像后,云服务器可将温度异常区域的可见光图像和温度异常区域的热红外图像,直接作为目标设备的温度异常区域的温度测量结果,反馈给工作人员,以使得工作人员基于该温度测量结果,找到目标设备的温度异常区域,并对目标设备的温度异常区域进行检修。
从上述方法可以看出:在工作人员的指令下,机器人可控制摄像头对包含目标设备和其余设备的目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,并发送至云服务器。接着,云
服务器可对目标区域的可见光图像和目标区域的热红外图像进行处理,从而得到目标设备的可见光图像和目标设备的热红外图像。然后,云服务器可继续对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像。最后,云服务器可基于温度异常区域的热红外图像,确定目标设备的温度异常区域的温度测量结果,并上报给工作人员,以使得工作人员对目标设备的温度异常区域进行检修。前述过程过程中,云服务器在获取目标区域的可见光图像和热红外图像后,可从中将其余设备的可见光图像和热红外图像剔除,得到目标设备的可见光图像和热红外图像。由于目标设备的可见光图像和热红外图像仅呈现目标设备,在对这些图像进行分析的过程中,云服务器不会受到其余设备的影响,精准确认目标设备中温度异常区域的可见光图像和热红外图像,以此来作为目标设备的温度异常区域的温度测量结果。由此可见,云服务器准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
在一种可能实现的方式中,该方法还包括:对目标区域的热红外图像以及目标区域的可见光图像进行计算,得到目标区域的深度图像;对目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像包括:对目标区域的深度图像、目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。前述实现方式中,得到目标区域的可见光图像以及目标区域的热红外图像后,云服务器可对目标区域的可见光图像以及目标区域的热红外图像进行一系列的计算,从而得到目标区域的深度图像。得到目标区域的深度图像后,云服务器可利用目标区域的深度图像,在目标区域的可见光图像和目标区域的热红外图像中,将目标区域中除目标设备之外的其余区域的可见光图像以及其余区域的热红外图像剔除,从而得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该方法还包括:将目标区域的可见光图像对齐至目标区域的热红外图像,得到目标区域的对齐后的可见光图像;对目标区域的热红外图像以及目标区域的可见光图像进行计算,得到目标区域的深度图像包括:对目标区域的热红外图像以及对齐后的可见光图像进行计算,得到目标区域的深度图像;对目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像包括:对目标区域的对齐后的可见光图像、目标区域的热红外图像和目标区域的深度图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。前述实现方式中,在得到目标区域的可见光图像以及目标区域的热红外图像后,云服务器可基于第一模型,对目标区域的可见光图像与目标区域的热红外图像进行对齐(包含粗糙化对应以及精细化对齐),从而得到目标区域的对齐后的可见光图像。在得到目标区域的对齐后的可见光图像后,云服务器还可基于第二模型,对目标区域的对齐后的可见光图像以及目标区域的热红外图像进行一系列的计算,从而得到目标区域的深度图像。在得到目标区域的深度图像后,云服务器可基于第三模型和第四模型,对目标区域的对齐后的可见光图像、目标区域的热红外图像和目标区域的深度图像完成实例分割,从而准确得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该方法还包括:基于目标区域的深度图像,在目标区域的热红外图像以及对齐后的可见光图像中,确定目标区域中前景的热红外图像以及前景的可见光图像,前景包含目标设备;对目标区域的对齐后的可见光图像、目标区域的热红外图像和目标区域的深度图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像包括:对目标区域的深度图像、前景的热红外图像以及前景的可见光图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。前述实现方式中,在得到目标区域的深度图像后,云服务器还可基于目标区域的深度图像,在目标区域的热红外图像以及目标区域的对齐后的可见光图像中,将目标区域中背景的可见光区域以及目标区域中背景的热红外图像剔除,保留目标区域中前景的热红外图像以及目标区域中前景的可见光图像,目标区域中前景的可见光图像仅呈现目标区域的前景,目标区域中前景的热红外图像仅呈现目标区域的前景,目标区域的前景包含目标设备以及其余设备,目标区域的背景为目标设备所处的环境。在得到目标区域中前景的热红外图像以及目标区域中前景的可见光图像后,云服务器可基于第三模型和第四模型,对目标区域中前景的热红外图像、目标区域中前景的可见光图像和目标区域的深度图像完成实例分割,从而准确得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,对目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标
设备的可见光图像和目标设备的热红外图像包括:对目标区域的可见光图像和目标区域的热红外图像进行分割,得到目标子区域的可见光图像和目标子区域的热红外图像,目标子区域为目标区域中,目标设备以及其余设备的一部分所占据的区域;对目标子区域的可见光图像和目标子区域的热红外图像进行二次分割,得到目标设备的可见光图像和目标设备的热红外图像。前述实现方式中,实例分割包含首次分割以及二次分割。在得到目标区域的深度图像后,云服务器可基于第三模型,利用目标区域的深度图像,对目标区域的热红外图像以及目标区域的可见光图像进行首次分割,从而得到目标子区域的热红外图像以及目标子区域的可见光图像。在得到目标子区域的可见光图像和目标子区域的热红外图像后,由于目标子区域为目标区域中目标设备以及其余设备的一部分所占据的区域,云服务器还可基于第四模型,对目标子区域的可见光图像和目标子区域的热红外图像进行二次分割,从而准确得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该方法还包括:基于目标区域的可见光图像以及目标区域的热红外图像,获取摄像头与温度异常区域之间的距离;基于摄像头与温度异常区域之间的距离和预设对应关系对温度测量结果进行调整,得到温度异常区域的调整后的温度测量结果,预设对应关系用于指示距离与温度修正值之间的对应关系。前述实现方式中,云服务器还设置有预设对应关系,该预设对应关系用于指示距离与温度修正值之间的对应关系。那么,云服务器在得到温度异常区域的热红外图像后,可基于该预设对应关系以及热成像摄像头与目标设备之间的距离(该距离可从目标区域的深度图像中获取),确定热成像摄像头与目标设备之间的距离所对应的温度修正值,再基于该温度修正值对温度异常区域的热红外图像进行调整,得到温度异常区域的调整后的热红外图像。如此一来,云服务器可将温度异常区域的可见光图像和温度异常区域的调整后的热红外图像,作为目标设备中温度异常区域的调整后的温度测量结果,并反馈给工作人员。由此可见,云服务器通过温度修正,可以令目标设备中温度异常区域的调整后的热红外图像更加贴近目标设备中温度异常区域的实际温度,有利于提高目标设备的温度测量结果的准确度。此外,由于目标区域的深度图像包含目标区域中各个点到摄像头之间的距离,云服务器可基于目标区域的深度图像自动获取摄像头与目标设备中温度异常区域之间的距离,以此来确定该距离对应的温度修正值,并利用温度修正值来完成温度测量结果的调整。由此可见,温度修正的整个过程可由云服务器自动完成,不需要工作人员进行操作,可降低人为操作成本。
在一种可能实现的方式中,控制摄像头对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像包括:控制摄像头在预置的位置按照预置的角度对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,位于位置上的摄像头与目标区域中的目标设备之间的距离位于预置范围内,摄像头按照角度对目标区域进行拍摄时,目标设备与其余设备的重叠程度小于预置阈值。前述实现方式中,在巡检规划阶段中,云服务器可基于机器人在不同位置以及不同角度所拍摄的目标区域的图像,按照某些条件为机器人挑选出预置的位置以及预置的角度,这些条件包含:机器人位于该位置上时,摄像头与目标区域中目标设备之间的距离位于预置范围内,且机器人控制摄像头按照该角度对目标区域进行拍摄时,目标设备与其余设备的重叠程度小于预置阈值。由此可见,云服务器依照前述的条件可自动为机器人规划出最优的位置和角度,所考虑的因素较为全面,这样可以使得机器人在巡检执行阶段中,在最优的位置和角度上拍摄出最优的图像,不仅效率高,还可提高目标设备的温度测量结果的准确度。
在一种可能实现的方式中,机器人的摄像头包含光成像摄像头以及热成像摄像头。
本申请实施例的第二方面提供了一种设备测温装置,该装置包括:拍摄模块,用于控制摄像头对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,目标区域包含目标设备;第一处理模块,用于对目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像;第二处理模块,用于对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像;确定模块,用于基于温度异常区域的热红外图像,确定温度异常区域的温度测量结果。
从上述装置可以看出:在工作人员的指令下,机器人可控制摄像头对包含目标设备和其余设备的目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,并发送至云服务器。接着,云服务器可对目标区域的可见光图像和目标区域的热红外图像进行处理,从而得到目标设备的可见光图像
和目标设备的热红外图像。然后,云服务器可继续对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像。最后,云服务器可基于温度异常区域的热红外图像,确定目标设备的温度异常区域的温度测量结果,并上报给工作人员,以使得工作人员对目标设备的温度异常区域进行检修。前述过程过程中,云服务器在获取目标区域的可见光图像和热红外图像后,可从中将其余设备的可见光图像和热红外图像剔除,得到目标设备的可见光图像和热红外图像。由于目标设备的可见光图像和热红外图像仅呈现目标设备,在对这些图像进行分析的过程中,云服务器不会受到其余设备的影响,精准确认目标设备中温度异常区域的可见光图像和热红外图像,以此来作为目标设备的温度异常区域的温度测量结果。由此可见,云服务器准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
在一种可能实现的方式中,该装置还包括:计算模块,用于对目标区域的热红外图像以及目标区域的可见光图像进行计算,得到目标区域的深度图像;第一处理模块,用于对目标区域的深度图像、目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该装置还包括:第三处理模块,用于将目标区域的可见光图像对齐至目标区域的热红外图像,得到目标区域的对齐后的可见光图像;计算模块,用于对目标区域的热红外图像以及对齐后的可见光图像进行计算,得到目标区域的深度图像;第一处理模块,用于对目标区域的对齐后的可见光图像、目标区域的热红外图像和目标区域的深度图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该装置还包括:第四处理模块,用于基于目标区域的深度图像,在目标区域的热红外图像以及对齐后的可见光图像中,确定目标区域中前景的热红外图像以及前景的可见光图像,前景包含目标设备;第一处理模块,用于对目标区域的深度图像、前景的热红外图像以及前景的可见光图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,第一处理模块,用于:对目标区域的可见光图像和目标区域的热红外图像进行分割,得到目标子区域的可见光图像和目标子区域的热红外图像,目标子区域为目标区域中,目标设备以及其余设备的一部分所占据的区域;对目标子区域的可见光图像和目标子区域的热红外图像进行二次分割,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该装置还包括:获取模块,用于基于所述目标区域的可见光图像以及所述目标区域的热红外图像,获取所述摄像头与所述温度异常区域之间的距离;调整模块,用于基于摄像头与温度异常区域之间的距离和预设对应关系对温度测量结果进行调整,得到温度异常区域的调整后的温度测量结果,预设对应关系用于指示距离与温度修正值之间的对应关系。
在一种可能实现的方式中,拍摄模块,用于控制摄像头在预置的位置按照预置的角度对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,位于位置上的摄像头与目标区域中的目标设备之间的距离位于预置范围内,摄像头按照角度对目标区域进行拍摄时,目标设备与其余设备的重叠程度小于预置阈值。
在一种可能实现的方式中,摄像头包含光成像摄像头以及热成像摄像头。
本申请实施例的第三方面提供了一种设备测温装置,该装置包括存储器和处理器;存储器存储有代码,处理器被配置为执行代码,当代码被执行时,设备测温装置执行如第一方面或第一方面中任意一种可能的实现方式所述的方法。
本申请实施例的第四方面提供了一种电路***,该电路***包括处理电路,该处理电路配置为执行如第一方面或第一方面中的任意一种可能的实现方式所述的方法。
本申请实施例的第五方面提供了一种芯片***,该芯片***包括处理器,用于调用存储器中存储的计算机程序或计算机指令,以使得该处理器执行如第一方面或第一方面中的任意一种可能的实现方式所述的方法。
在一种可能的实现方式中,该处理器通过接口与存储器耦合。
在一种可能的实现方式中,该芯片***还包括存储器,该存储器中存储有计算机程序或计算机指令。
本申请实施例的第六方面提供了一种计算机存储介质,该计算机存储介质存储有计算机程序,该程
序在由计算机执行时,使得计算机实施如第一方面或第一方面中的任意一种可能的实现方式所述的方法。
本申请实施例的第七方面提供了一种计算机程序产品,该计算机程序产品存储有指令,该指令在由计算机执行时,使得计算机实施如第一方面或第一方面中的任意一种可能的实现方式所述的方法。
本申请实施例中,在工作人员的指令下,机器人可控制摄像头对包含目标设备和其余设备的目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,并发送至云服务器。接着,云服务器可对目标区域的可见光图像和目标区域的热红外图像进行处理,从而得到目标设备的可见光图像和目标设备的热红外图像。然后,云服务器可继续对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像。最后,云服务器可基于温度异常区域的热红外图像,确定目标设备的温度异常区域的温度测量结果,并上报给工作人员,以使得工作人员对目标设备的温度异常区域进行检修。前述过程过程中,云服务器在获取目标区域的可见光图像和热红外图像后,可从中将其余设备的可见光图像和热红外图像剔除,得到目标设备的可见光图像和热红外图像。由于目标设备的可见光图像和热红外图像仅呈现目标设备,在对这些图像进行分析的过程中,云服务器不会受到其余设备的影响,精准确认目标设备中温度异常区域的可见光图像和热红外图像,以此来作为目标设备的温度异常区域的温度测量结果。由此可见,云服务器准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
图1为人工智能主体框架的一种结构示意图;
图2a为本申请实施例提供的设备测温***的一个结构示意图;
图2b为本申请实施例提供的设备测温***的另一结构示意图;
图2c为本申请实施例提供的设备测温的相关设备的一个示意图;
图3为本申请实施例提供的***100架构的一个示意图;
图4为本申请实施例提供的设备测温方法的一个流程示意图;
图5为本申请实施例提供的校准标定板的一个结构示意图;
图6a为本申请实施例提供的标定过程的一个示意图;
图6b为本申请实施例提供的标定过程的另一示意图;
图6c为本申请实施例提供的标定过程的另一示意图;
图7为本申请实施例提供的精细化匹配的一个示意图;
图8为本申请实施例提供的第一模型的一个结构示意图;
图9为本申请实施例提供的第二模型的一个结构示意图;
图10为本申请实施例提供的异源双目视差的一个示意图;
图11为本申请实施例提供的第三模型的一个结构示意图;
图12为本申请实施例提供的第四模型的一个结构示意图;
图13为本申请实施例提供的第五模型的一个结构示意图;
图14为本申请实施例提供的设备测温装置的一个结构示意图;
图15为本申请实施例提供的执行设备的一个结构示意图;
图16为本申请实施例提供的训练设备的一个结构示意图;
图17为本申请实施例提供的芯片的一个结构示意图。
本申请实施例提供了一种设备测温方法及其相关设备,可准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象,而不必用于描述特定的顺序或先后次序。应该理解这样使用的术语在适当情况下可以互换,这仅仅是描述本申请的实施例中对相同属性的对象在描述时所采用的区分方式。此外,术语“包括”和“具有”以及他们的任何变形,意图在于覆盖不排他的包含,以便包含一系列单元的过程、方法、***、产品或设
备不必限于那些单元,而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它单元。
变电站中电力设备的预防性检测是维护电力设备的重要手段,测量电力设备的温度又是预防性检测中的重要环节,通过测量电力设备的温度可以确定电力设备中的温度正常区域以及温度异常区域,从而判断电力设备是否能够正常工作,即电力设备是否存在故障。
目前,变电站中的机器人在巡检时,可拍摄包含目标电力设备的某一个区域的热红外图像,并将该图像发送至远端的云服务器进行图像分析。云服务器基于该图像可确定目标电力设备的热红外图像,也就是目标电力设备所有点的温度,故云服务器可确定目标电力设备的温度正常区域以及温度异常区域,以此作为目标电力设备的温度测量结果并及时通知变电站的工作人员,以使得工作人员对目标电力设备的温度异常区域进行检修。
然而,变电站中的电力设备很密集,该区域有可能不仅包含目标电力设备,还包含其余电力设备,云服务器在对该区域的热红外图像进行分析时,容易受到其余电力设备的影响,将其余电力设备边缘附近的某些点的温度误认为目标电力设备中某些点的温度,误判了目标电力设备的温度异常区域,导致目标电力设备的温度测量结果的准确度较低。
进一步地,机器人在巡检过程中,往往需要工作人员为机器人来设定某个位置,以使得机器人停靠在该位置上,对包含目标电力设备的该区域进行拍摄,人为选定位置往往考虑的因素较为单一,导致拍摄得到的该区域的热红外图像不是最优的图像,不仅效率低下,还会降低目标电力设备的温度测量结果的准确度。
更进一步地,机器人采用热成像摄像头来获取该区域的热红外图像,图像中任意一个点的温度(即图像中任意一个像素)均是由热成像摄像头基于接收到的辐射能量计算得出,由于辐射能量受到目标电力设备和摄像头之间的距离的影响,会导致目标电力设备上各个点的测量温度与实际温度存在偏差,也会降低目标电力设备的温度测量结果的准确度。
为了解决上述问题,本申请实施例提供了一种设备测温方法,该方法可结合人工智能(artificial intelligence,AI)技术实现。AI技术是利用数字计算机或者数字计算机控制的机器模拟、延伸和扩展人的智能的技术学科,AI技术通过感知环境、获取知识并使用知识获得最佳结果。换句话说,人工智能技术是计算机科学的一个分支,它企图了解智能的实质,并生产出一种新的能以人类智能相似的方式做出反应的智能机器。利用人工智能进行数据处理是人工智能常见的一个应用方式。
首先对人工智能***总体工作流程进行描述,请参见图1,图1为人工智能主体框架的一种结构示意图,下面从“智能信息链”(水平轴)和“IT价值链”(垂直轴)两个维度对上述人工智能主题框架进行阐述。其中,“智能信息链”反映从数据的获取到处理的一列过程。举例来说,可以是智能信息感知、智能信息表示与形成、智能推理、智能决策、智能执行与输出的一般过程。在这个过程中,数据经历了“数据—信息—知识—智慧”的凝练过程。“IT价值链”从人智能的底层基础设施、信息(提供和处理技术实现)到***的产业生态过程,反映人工智能为信息技术产业带来的价值。
(1)基础设施
基础设施为人工智能***提供计算能力支持,实现与外部世界的沟通,并通过基础平台实现支撑。通过传感器与外部沟通;计算能力由智能芯片(CPU、NPU、GPU、ASIC、FPGA等硬件加速芯片)提供;基础平台包括分布式计算框架及网络等相关的平台保障和支持,可以包括云存储和计算、互联互通网络等。举例来说,传感器和外部沟通获取数据,这些数据提供给基础平台提供的分布式计算***中的智能芯片进行计算。
(2)数据
基础设施的上一层的数据用于表示人工智能领域的数据来源。数据涉及到图形、图像、语音、文本,还涉及到传统设备的物联网数据,包括已有***的业务数据以及力、位移、液位、温度、湿度等感知数据。
(3)数据处理
数据处理通常包括数据训练,机器学习,深度学习,搜索,推理,决策等方式。
其中,机器学习和深度学习可以对数据进行符号化和形式化的智能信息建模、抽取、预处理、训练
等。
推理是指在计算机或智能***中,模拟人类的智能推理方式,依据推理控制策略,利用形式化的信息进行机器思维和求解问题的过程,典型的功能是搜索与匹配。
决策是指智能信息经过推理后进行决策的过程,通常提供分类、排序、预测等功能。
(4)通用能力
对数据经过上面提到的数据处理后,进一步基于数据处理的结果可以形成一些通用的能力,比如可以是算法或者一个通用***,例如,翻译,文本的分析,计算机视觉的处理,语音识别,图像的识别等等。
(5)智能产品及行业应用
智能产品及行业应用指人工智能***在各领域的产品和应用,是对人工智能整体解决方案的封装,将智能信息决策产品化、实现落地应用,其应用领域主要包括:智能终端、智能交通、智能医疗、自动驾驶、智慧城市等。
接下来介绍几种本申请的应用场景。
图2a为本申请实施例提供的设备测温***的一个结构示意图,该设备测温***包括用户设备以及数据处理设备。其中,用户设备包括手机、个人电脑或者机器人等智能终端。用户设备为设备测温的发起端,作为设备测温请求的发起方,通常由用户通过用户设备发起请求。
上述数据处理设备可以是云服务器、网络服务器、应用服务器以及管理服务器等具有数据处理功能的设备或服务器。数据处理设备通过交互接口接收来自智能终端的设备测温请求,再通过存储数据的存储器以及数据处理的处理器环节进行机器学习,深度学习,搜索,推理,决策等方式的文本处理。数据处理设备中的存储器可以是一个统称,包括本地存储以及存储历史数据的数据库,数据库可以在数据处理设备上,也可以在其它网络服务器上。
在图2a所示的设备测温***中,用户设备可以接收用户的指令,用户设备可基于该指令,对位于目标区域中的目标设备进行拍摄,从而得到目标区域的图像,然后向数据处理设备发起温度测量请求,使得数据处理设备基于该请求针对用户设备得到的图像执行图像处理应用,从而得到图像的处理结果。示例性的,用户设备在用户的指令下,采集目标区域的可见光图像以及目标区域的热红外图像,并向数据处理设备发送针对这些图像的温度测量请求,以使得数据处理设备对目标区域的可见光图像以及目标区域的热红外图像进行一系列的分析处理,从而得到图像的处理结果,即目标设备的温度测量结果。
在图2a中,数据处理设备可以执行本申请实施例的设备测温方法。
图2b为本申请实施例提供的设备测温***的另一结构示意图,在图2b中,用户设备直接作为数据处理设备,该用户设备能够在用户的指令下,获取来自目标区域的图像并直接由用户设备本身的硬件进行图像处理应用,具体过程与图2a相似,可参考上面的描述,在此不再赘述。
在图2b所示的设备测温***中,用户设备在用户的指令下,采集目标区域的可见光图像以及目标区域的热红外图像,并对目标区域的可见光图像以及目标区域的热红外图像进行一系列的分析处理,从而得到图像的处理结果,目标设备的温度测量结果。
在图2b中,用户设备自身就可以执行本申请实施例的设备测温方法。
图2c为本申请实施例提供的设备测温的相关设备的一个示意图。
上述图2a和图2b中的用户设备具体可以是图2c中的本地设备301或者本地设备302,图2a中的数据处理设备具体可以是图2c中的执行设备210,其中,数据存储***250可以存储执行设备210的待处理数据,数据存储***250可以集成在执行设备210上,也可以设置在云上或其它网络服务器上。
图2a和图2b中的处理器可以通过神经网络模型或者其它模型(例如,基于支持向量机的模型)进行数据训练/机器学习/深度学习,并利用数据最终训练或者学习得到的模型针对图像执行图像处理应用,从而得到相应的处理结果。
图3为本申请实施例提供的***100架构的一个示意图,在图3中,执行设备110配置输入/输出(input/output,I/O)接口112,用于与外部设备进行数据交互,用户可以通过客户设备140向I/O接口112输入数据,所述输入数据在本申请实施例中可以包括:温度测量任务、任务数据(包含待处理的可见光图像和热红外图像)以及任务参数。
在执行设备110对输入数据进行预处理,或者在执行设备110的计算模块111执行计算等相关的处理(比如进行本申请中神经网络的功能实现)过程中,执行设备110可以调用数据存储***150中的数据、代码等以用于相应的处理,也可以将相应处理得到的数据、指令等存入数据存储***150中。
最后,I/O接口112将处理结果返回给客户设备140,从而提供给用户。
值得说明的是,训练设备120可以针对不同的目标或称不同的任务,基于不同的训练数据生成相应的目标模型/规则,该相应的目标模型/规则即可以用于实现上述目标或完成上述任务,从而为用户提供所需的结果。其中,训练数据可以存储在数据库130中,且来自于数据采集设备160采集的训练样本。此外,训练设备120与执行设备110既可以是不同的设备,也可以是同一个设备(即训练设备120集成在执行设备110中)。
在图3中所示情况下,用户可以手动给定输入数据,该手动给定可以通过I/O接口112提供的界面进行操作。另一种情况下,客户设备140可以自动地向I/O接口112发送输入数据,如果要求客户设备140自动发送输入数据需要获得用户的授权,则用户可以在客户设备140中设置相应权限。用户可以在客户设备140查看执行设备110输出的结果,具体的呈现形式可以是显示、声音、动作等具体方式。客户设备140也可以作为数据采集端,采集如图所示输入I/O接口112的输入数据及输出I/O接口112的输出结果作为新的样本数据,并存入数据库130。当然,也可以不经过客户设备140进行采集,而是由I/O接口112直接将如图所示输入I/O接口112的输入数据及输出I/O接口112的输出结果,作为新的样本数据存入数据库130。
值得注意的是,图3仅是本申请实施例提供的一种***架构的示意图,图中所示设备、器件、模块等之间的位置关系不构成任何限制,例如,在图3中,数据存储***150相对执行设备110是外部存储器,在其它情况下,也可以将数据存储***150置于执行设备110中。如图3所示,可以根据训练设备120训练得到神经网络。
本申请实施例还提供的一种芯片,该芯片包括神经网络处理器NPU。该芯片可以被设置在如图3所示的执行设备110中,用以完成计算模块111的计算工作。该芯片也可以被设置在如图3所示的训练设备120中,用以完成训练设备120的训练工作并输出目标模型/规则。
神经网络处理器NPU,NPU作为协处理器挂载到主中央处理器(central processing unit,CPU)(host CPU)上,由主CPU分配任务。NPU的核心部分为运算电路,控制器控制运算电路提取存储器(权重存储器或输入存储器)中的数据并进行运算。
在一些实现中,运算电路内部包括多个处理单元(process engine,PE)。在一些实现中,运算电路是二维脉动阵列。运算电路还可以是一维脉动阵列或者能够执行例如乘法和加法这样的数学运算的其它电子线路。在一些实现中,运算电路是通用的矩阵处理器。
举例来说,假设有输入矩阵A,权重矩阵B,输出矩阵C。运算电路从权重存储器中取矩阵B相应的数据,并缓存在运算电路中每一个PE上。运算电路从输入存储器中取矩阵A数据与矩阵B进行矩阵运算,得到的矩阵的部分结果或最终结果,保存在累加器(accumulator)中。
向量计算单元可以对运算电路的输出做进一步处理,如向量乘,向量加,指数运算,对数运算,大小比较等等。例如,向量计算单元可以用于神经网络中非卷积/非FC层的网络计算,如池化(pooling),批归一化(batch normalization),局部响应归一化(local response normalization)等。
在一些实现中,向量计算单元能将经处理的输出的向量存储到统一缓存器。例如,向量计算单元可以将非线性函数应用到运算电路的输出,例如累加值的向量,用以生成激活值。在一些实现中,向量计算单元生成归一化的值、合并值,或二者均有。在一些实现中,处理过的输出的向量能够用作到运算电路的激活输入,例如用于在神经网络中的后续层中的使用。
统一存储器用于存放输入数据以及输出数据。
权重数据直接通过存储单元访问控制器(direct memory access controller,DMAC)将外部存储器中的输入数据搬运到输入存储器和/或统一存储器、将外部存储器中的权重数据存入权重存储器,以及将统一存储器中的数据存入外部存储器。
总线接口单元(bus interface unit,BIU),用于通过总线实现主CPU、DMAC和取指存储器之间进行交互。
与控制器连接的取指存储器(instruction fetch buffer),用于存储控制器使用的指令;
控制器,用于调用指存储器中缓存的指令,实现控制该运算加速器的工作过程。
一般地,统一存储器,输入存储器,权重存储器以及取指存储器均为片上(On-Chip)存储器,外部存储器为该NPU外部的存储器,该外部存储器可以为双倍数据率同步动态随机存储器(double data rate synchronous dynamic random access memory,DDR SDRAM)、高带宽存储器(high bandwidth memory,HBM)或其他可读可写的存储器。
由于本申请实施例涉及大量神经网络的应用,为了便于理解,下面先对本申请实施例涉及的相关术语及神经网络等相关概念进行介绍。
(1)神经网络
神经网络可以是由神经单元组成的,神经单元可以是指以xs和截距1为输入的运算单元,该运算单元的输出可以为:
其中,s=1、2、……n,n为大于1的自然数,Ws为xs的权重,b为神经单元的偏置。f为神经单元的激活函数(activation functions),用于将非线性特性引入神经网络中,来将神经单元中的输入信号转换为输出信号。该激活函数的输出信号可以作为下一层卷积层的输入。激活函数可以是sigmoid函数。神经网络是将许多个上述单一的神经单元联结在一起形成的网络,即一个神经单元的输出可以是另一个神经单元的输入。每个神经单元的输入可以与前一层的局部接受域相连,来提取局部接受域的特征,局部接受域可以是由若干个神经单元组成的区域。
神经网络中的每一层的工作可以用数学表达式y=a(Wx+b)来描述:从物理层面神经网络中的每一层的工作可以理解为通过五种对输入空间(输入向量的集合)的操作,完成输入空间到输出空间的变换(即矩阵的行空间到列空间),这五种操作包括:1、升维/降维;2、放大/缩小;3、旋转;4、平移;5、“弯曲”。其中1、2、3的操作由Wx完成,4的操作由+b完成,5的操作则由a()来实现。这里之所以用“空间”二字来表述是因为被分类的对象并不是单个事物,而是一类事物,空间是指这类事物所有个体的集合。其中,W是权重向量,该向量中的每一个值表示该层神经网络中的一个神经元的权重值。该向量W决定着上文所述的输入空间到输出空间的空间变换,即每一层的权重W控制着如何变换空间。训练神经网络的目的,也就是最终得到训练好的神经网络的所有层的权重矩阵(由很多层的向量W形成的权重矩阵)。因此,神经网络的训练过程本质上就是学习控制空间变换的方式,更具体的就是学习权重矩阵。
因为希望神经网络的输出尽可能地接近真正想要预测的值,所以可以通过比较当前网络的预测值和真正想要的目标值,再根据两者之间的差异情况来更新每一层神经网络的权重向量(当然,在第一次更新之前通常会有初始化的过程,即为神经网络中的各层预先配置参数),比如,如果网络的预测值高了,就调整权重向量让它预测低一些,不断地调整,直到神经网络能够预测出真正想要的目标值。因此,就需要预先定义“如何比较预测值和目标值之间的差异”,这便是损失函数(loss function)或目标函数(objective function),它们是用于衡量预测值和目标值的差异的重要方程。其中,以损失函数举例,损失函数的输出值(loss)越高表示差异越大,那么神经网络的训练就变成了尽可能缩小这个loss的过程。
(2)反向传播算法
神经网络可以采用误差反向传播(back propagation,BP)算法在训练过程中修正初始的神经网络模型中参数的大小,使得神经网络模型的重建误差损失越来越小。具体地,前向传递输入信号直至输出会产生误差损失,通过反向传播误差损失信息来更新初始的神经网络模型中参数,从而使误差损失收敛。反向传播算法是以误差损失为主导的反向传播运动,旨在得到最优的神经网络模型的参数,例如权重矩阵。
下面将结合神经网络的训练侧和神经网络的应用侧对本申请提供的方法进行描述。
本申请实施例提供的模型训练方法,涉及数据序列的处理,具体可以应用于数据训练、机器学习、深度学习等方法,对训练数据进行符号化和形式化的智能信息建模、抽取、预处理、训练等,最终得到训练好的神经网络(如本申请中的第一模型、第二模型、第三模型、第四模型以及第五模型等等);并且,本申请实施例提供的设备测温方法可以运用上述训练好的神经网络,将输入数据(例如,将本申请中的目标区域的可见光图像以及目标区域的热红外图像等等)输入到所述训练好的神经网络中,得到输出数据(如本申请中目标设备的温度异常区域的可见光图像以及温度异常区域的热红外图像等等)。需要说明的是,本申请实施例提供的模型训练方法和设备测温方法是基于同一个构思产生的发明,也可以理解为一个***中的两个部分,或一个整体流程的两个阶段:如模型训练阶段和模型应用阶段。
需要说明的是,本申请实施例提供的设备测温方法可应用于电力设备测温场景中,在该场景中,变电站的工作人员可令机器人在变电站中巡检。在连续行驶过程中,机器人控制摄像头对目标电力设备以及其余电力设备所在的目标区域进行拍摄,并将得到的目标区域的可见光图像以及热红外图像上传至云服务器,以使得云服务器对这些图像进行一系列的处理,得到目标电力设备中温度异常区域的温度测量结果,并反馈给工作人员,以使得工作人员在变电站中,对目标电力设备的温度异常区域进行检修,从而维护变电站的正常工作状态。图4为本申请实施例提供的设备测温方法的一个流程示意图,如图4所示,该方法包括:
401、控制摄像头对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,目标区域包含目标设备。
本实施例中,工作人员对机器人下发指令,机器人基于该指令可确定需要完成的巡检任务,即完成针对目标设备的温度测量。那么,在巡检执行阶段(即执行巡检任务的阶段)中,机器人可控制摄像头对目标设备所在的目标区域进行拍摄,从而得到目标区域的可见光图像以及目标区域的热红外图像,可以理解的是,目标区域通常指被圈定某个范围的区域,该区域内包含目标设备以及区域设备,目标区域的可见光图像包含目标区域中各个点的原始信息(即各个点颜色、纹理等等),目标区域的热红外图像包含目标区域中各个点的温度信息(即各个点的温度)。
具体地,设置在机器人上的摄像头可包含光成像摄像头以及热成像摄像头,且光成像摄像头以及热成像摄像头并排设置在机器人的云台上,云台可在机器人的控制下,实现360度的旋转。那么,机器人可在连续行驶的过程中,控制光成像摄像头对目标区域进行拍摄,从而得到目标区域的可见光图像,并控制热成像摄像头对目标区域进行拍摄,从而得到目标区域的热红外图像。
更具体地,机器人可通过以下方式对目标区域进行拍摄,从而得到目标区域的可见光图像和目标区域的热红外图像包括:
(1)在连续行驶的过程中,机器人到达预置的位置时(并不用停下来),可在这一瞬间立即控制摄像头按照预置的角度对目标区域进行拍摄,从而得到目标区域的可见光图像和目标区域的热红外图像。需要说明的是,预置的位置以及预置的角度均是巡检规划阶段中提前确定的,那么,提前确定的该位置应满足以下条件:机器人位于该位置上时,机器人的摄像头与目标区域中的目标设备之间的距离位于预置范围内(该范围的左右端值的大小可根据实际需求来设置,此处不做限制)。提前确定的该角度满足以下条件:机器人的摄像头按照该角度对目标区域进行拍摄时,所呈现的拍摄视野中,目标区域中的目标设备与目标区域中的其余设备之间的重叠程度小于预置的重叠程度阈值(该阈值的大小可根据实际需求来设置,此处不做限制)。
需要说明的是,巡检规划阶段位于巡检执行阶段之前,且巡检规划阶段与巡检执行阶段是类似的,故机器人完成巡检规划阶段的过程可参考机器人完成巡检执行阶段的过程(即参考步骤401以及步骤402等等),此处不做赘述。
(2)在连续行驶的过程中,机器人每到达一个位置,则选择一个角度,并在这一瞬间立即判断该位置应满足以下条件:机器人位于该位置上时,机器人的摄像头与目标区域中的目标设备之间的距离是否位于预置范围内,且判断某个角度是否满足以下条件:机器人的摄像头按照该角度对目标区域进行拍摄时,所呈现的拍摄视野中,目标区域中的目标设备与目标区域中的其余设备之间的重叠程度是否小于预置阈值。若这两个条件均满足,机器人则在该位置,控制摄像头按照该角度对目标区域进行拍摄,从而得到目标区域的可见光图像和目标区域的热红外图像,若这两个条件有一个不满足,机器人则在下一
个位置以及下一个角度,重新执行前述过程,直至成功采集到目标区域的可见光图像和目标区域的热红外图像。
402、对目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
得到目标区域的可见光图像以及目标区域的热红外图像后,机器人可将目标区域的可见光图像以及目标区域的热红外图像发送至云服务器,云服务器可在目标区域的可见光图像和目标区域的热红外图像中,将目标区域中除目标设备之外的其余区域的可见光图像以及其余区域的热红外图像剔除,从而得到目标设备的可见光图像和目标设备的热红外图像。可以理解的是,目标区域的可见光图像呈现了整个目标区域,目标区域的热红外图像呈现了整个目标区域,目标设备的可见光图像仅呈现目标设备,目标设备的热红外图像仅呈现目标设备。
具体地,云服务器可通过以下方式来获取目标设备的可见光图像和目标设备的热红外图像,该过程包括:
(1)在得到目标区域的可见光图像以及目标区域的热红外图像后,由于光成像摄像头的视野范围通常比热成像摄像头的视野范围大,导致目标区域的可见光图像的尺寸大于目标区域的热红外图像的尺寸,即前者所呈现的内容多于后者所呈现的内容,为了令二者所呈现的内容尽可能相同且对齐,云服务器可将目标区域的可见光图像对齐至目标区域的热红外图像,得到目标区域的对齐后的可见光图像。值得注意的是,该对齐过程包含初次匹配(粗糙化匹配)以及二次匹配(精细化匹配)两个阶段,下文分别对这两个阶段进行介绍:
(1.1)初次匹配阶段:在得到目标区域的可见光图像以及目标区域的热红外图像后,云服务器可基于可见光图像与热红外图像之间的内容匹配关系,对目标区域的可见光图像进行裁剪,以使得目标区域的裁剪后的可见光图像的尺寸与目标区域的热红外图像的尺寸大致相似(一般地,目标区域的裁剪后的可见光图像的尺寸稍微大于目标区域的热红外图像的尺寸)。此时,目标区域的裁剪后的可见光图像所呈现的内容与目标区域的热红外图像所呈现的内容大致相似(一般地,目标区域的裁剪后的可见光图像所呈现的内容稍微多于目标区域的热红外图像所呈现的内容)。
需要说明的是,可见光图像与热红外图像之间的对应关系是提前在巡检规划阶段中确定的。如图5所示(图5为本申请实施例提供的校准标定板的一个结构示意图),在巡检规划阶段中,工作人员可手持校准标定板,校准标定板包含依次层叠的标定板支撑面板、加热面板以及镂空圆形点阵玻璃面板,其中,镂空圆形点阵玻璃面板位于顶层,标定板支撑面板位于底层,位于中间层的加热面板可通过温度控制器来进行加热。工作人员可控制机器人通过摄像头对校准标定板进行拍摄,从而得到校准标定板的可见光图像以及校准标定板的热红外图像,并上传到云服务器。如图6a、图6b以及图6c所示(图6a为本申请实施例提供的标定过程的一个示意图、图6b为本申请实施例提供的标定过程的另一示意图、图6c为本申请实施例提供的标定过程的另一示意图),由于校准标定板的表面具有镂空圆形点阵(包含多个镂空圆形点),故云服务器可确定多个镂空圆形点在校准标定板的可见光图像中的位置,以及在校准标定板的热红外图像中的位置,并基于多个镂空圆形点在这两个图像中的位置,可确定校准标定板的可见光图像中哪一部分图像无法在校准标定板的热红外图像中找到相匹配的图像(即可见光图像中的这一部分图像所呈现的内容,在热红外图像所呈现的内容中并未出现),那么,校准标定板的可见光图像中的这一部分图像可被裁剪,从而得到校准标定板的裁剪后的可见光图像。如此一来,云服务器可记录被裁剪的这一部分图像在校准标定板的可见光图像中的位置,该位置信息可视为可见光图像与热红外图像之间的内容匹配关系。
(1.2)二次匹配阶段:得到目标区域的裁剪后的可见光图像后,云服务器可将目标区域的裁剪后的可见光图像以及目标区域的热红外图像输入至第一模型。如图7所示(图7为本申请实施例提供的精细化匹配的一个示意图),第一模型可计算目标区域的裁剪后的可见光图像中某一个像素点与目标区域的热红外图像中所有像素点之间距离,并将热红外图像中距离最小的像素点作为与可见光图像中该像素点相匹配的像素点。对于目标区域的裁剪后的可见光图像中的部分其余像素点,第一模型也可以执行如同对该像素点所执行的操作,故第一模型可得到目标区域的裁剪后的可见光图像与目标区域的热红外图像之间的像素点匹配关系。基于像素点匹配关系,第一模型可将目标区域的裁剪后的可见光图像的所有
像素点匹配至目标区域的热红外图像的所有像素点,即可将目标区域的裁剪后的可见光图像准确投影在目标区域的热红外图像上,完成投影的可见光图像即为目标区域的对齐后的可见光图像。
需要说明的是,前述的第一模型为已训练的神经网络模型,其结构如图8所示(图8为本申请实施例提供的第一模型的一个结构示意图),第一模型包含:VGG网络模块、位置信息编码模块、注意力模块、上采样模块、特征处理模块以及投影变换模块。
还需要说明的是,为了训练得到第一模型,云服务器可先获取第一待训练模型(即需要训练的神经网络模型)以及一批训练数据(包含某个区域的裁剪后的可见光图像以及该区域的热红外图像),训练数据的真实处理结果(即该区域的对齐后的真实可见光图像)。然后,云服务器可将训练数据输入至第一待训练模型,以通过第一待训练模型对训练数据进行处理,得到训练数据的预估处理结果(即该区域的对齐后的预估可见光图像)。随后,云服务器可通过预置的第一损失函数对训练数据的真实处理结果以及训练数据的预估处理结果进行计算,得到目标损失,目标损失用于指示训练数据的真实处理结果以及训练数据的预估处理结果之间的差异。最后,云服务器可基于目标损失更新第一待训练模型的参数,并利用下一批训练数据对更新参数后的第一待训练模型继续进行训练,直至满足模型训练条件(例如,目标损失收敛等等),从而得到第一模型。
(2)在得到目标区域的对齐后的可见光图像后,云服务器还对目标区域的热红外图像以及对齐后的可见光图像进行计算,得到目标区域的深度图像,目标区域的深度图像包含目标区域中各个点的深度信息(即各个点的深度等等)。该计算过程包括:
(2.1)在得到目标区域的对齐后的可见光图像后,云服务器可将目标区域的对齐后的可见光图像转化为目标区域的伪红外图像,并将目标区域的伪红外图像和目标区域的热红外图像输入至第二模型,以通过第二模型可对这些图像进行一系列的计算,从而得到视差图,视差图包含伪红外图像中每个像素点与热红外图像中相应的像素点之间的视差。
需要说明的是,前述的第二模型为已训练的神经网络模型,其结构如图9所示(图9为本申请实施例提供的第二模型的一个结构示意图),第二模型可包含:伪红外图像个体特征提取模块、伪红外图像与热红外图像共同特征提取模块、热红外图像个体特征提取模块、特征融合模块以及高分辨率重建模块。
还需要说明的是,为了训练得到第二模型,云服务器可先获取第二待训练模型(即需要训练的神经网络模型)以及一批训练数据(包含某个区域的伪红外图像以及该区域的热红外图像),训练数据的真实处理结果(即真实视差图)。然后,云服务器可将训练数据输入至第二待训练模型,以通过第二待训练模型对训练数据进行处理,得到训练数据的预估处理结果(即预估视差图)。随后,云服务器可通过预置的第二损失函数对训练数据的真实处理结果以及训练数据的预估处理结果进行计算,得到目标损失,目标损失用于指示训练数据的真实处理结果以及训练数据的预估处理结果之间的差异。最后,云服务器可基于目标损失更新第二待训练模型的参数,并利用下一批训练数据对更新参数后的第二待训练模型继续进行训练,直至满足模型训练条件(例如,目标损失收敛等等),从而得到第二模型。
(2.2)得到视差图后,对于伪红外图像的任意一个像素点,云服务器可对伪红外图像中该像素点的坐标、伪红外图像中该像素点与热红外图像中相应的像素点之间的视差进行计算,从而得到热红外图像中相应的像素点的坐标。然后,云服务器可对摄像头的内外参、该像素点的坐标、相应的像素点的坐标进行计算,确定该像素点对应的深度。对于伪红外图像的其余像素点,云服务器也可对其执行如同对该像素点所执行的操作,从而得到伪红外图像的所有像素点对应的深度,即目标区域中所有点的深度,目标区域中所有点的深度则构成了目标区域的深度图像。
例如,如图10所示(图10为本申请实施例提供的异源双目视差的一个示意图),云服务器可获取光成像摄像头的光心坐标Oir,热成像摄像头的光心坐标Orgb,两个光心坐标之间的差值为T,光成像摄像头的焦距fir,热成像摄像头的焦距frgb。对于目标区域的伪红外图像中的某个像素点,设该像素点的坐标为Xrgb,该像素点在目标区域的热红外图像中相应的像素点的坐标为Xir,Xrgb与Xir之间的差值为伪红外图像中该像素点与热红外图像中相应的像素点之间的视差。那么,该像素点对应的深度Z可通过以下公式计算得到:
如此一来,可计算所有像素点对应的深度,得到目标区域的深度图像。
(3)在得到目标区域的深度图像后,云服务器基于目标区域的深度图像,可在目标区域的热红外图像以及目标区域的对齐后的可见光图像中,确定目标区域中前景的热红外图像以及目标区域中前景的可见光图像,目标区域中前景的可见光图像仅呈现目标区域的前景,目标区域中前景的热红外图像仅呈现目标区域的前景,目标区域的前景包含目标设备以及其余设备,目标区域的背景为目标设备所处的环境(例如,远处的天空、山峰等等)。该确定过程包括:
在得到目标区域的深度图像后,云服务器基于目标区域的深度图像,可确定目标区域的热红外图像中所有像素点对应的深度,以及目标区域的对齐后的可见光图像中所有像素点对应的深度。在目标区域的热红外图像中,云服务器可将深度大于预置的深度阈值(该阈值的大小可根据实际需求进行设置,此处不做限制)的像素点剔除,相当于将目标区域中背景的热红外图像剔除,保留目标区域中前景的热红外图像。同理,在目标区域的可见光图像中,云服务器可将深度大于预置的深度阈值的像素点剔除,相当于将目标区域中背景的可见光图像剔除,保留目标区域中前景的可见光图像。
(4)在得到目标区域的深度图像后,云服务器可基于目标区域的深度图像,对目标区域中前景的热红外图像以及目标区域中前景的可见光图像进行分割,得到目标子区域的可见光图像和目标子区域的热红外图像,目标子区域(也可以称为目标检测框)为目标区域中,目标设备以及其余设备的一部分所占据的区域。该分割过程包括:
在得到目标区域的深度图像后,云服务器可将目标区域的深度图像、目标区域中前景的热红外图像以及目标区域中前景的可见光图像输入至第三模型,以使得第三模型基于目标区域的深度图像,在目标区域中前景的热红外图像中,将前景中除目标子区域之外的其余区域的热红外图像剔除,保留目标子区域的热红外图像。同理,第三模型还可在目标区域中前景的可见光图像中,将前景中除目标子区域之外的其余区域的可见光图像剔除,保留目标子区域的可见光图像。
需要说明的是,前述的第三模型为已训练的神经网络模型,其结构如图11所示(图11为本申请实施例提供的第三模型的一个结构示意图),第三模型可包含:可见光图像特征编码器、热红外图像特征编码器、特征融合模块、基于双空间图的特征融合模块、解码器、卷积核(1*1卷积核以及3*3卷积核等等)以及上采样模块。
还需要说明的是,为了训练得到第三模型,云服务器可先获取第三待训练模型(即需要训练的神经网络模型)以及一批训练数据(包含某个区域的深度图像、该区域中前景的热红外图像以及该区域中前
景的可见光图像),训练数据的真实处理结果(即检测框的真实可见光图像以及真实热红外图像)。然后,云服务器可将训练数据输入至第三待训练模型,以通过第三待训练模型对训练数据进行处理,得到训练数据的预估处理结果(即检测框的预估可见光图像以及预估热红外图像)。随后,云服务器可通过预置的第三损失函数对训练数据的真实处理结果以及训练数据的预估处理结果进行计算,得到目标损失,目标损失用于指示训练数据的真实处理结果以及训练数据的预估处理结果之间的差异。最后,云服务器可基于目标损失更新第三待训练模型的参数,并利用下一批训练数据对更新参数后的第三待训练模型继续进行训练,直至满足模型训练条件(例如,目标损失收敛等等),从而得到第三模型。
(5)在得到目标子区域的可见光图像和目标子区域的热红外图像后,云服务器可对目标子区域的可见光图像和目标子区域的热红外图像进行二次分割,从而得到目标设备的可见光图像和目标设备的热红外图像。该二次分割过程包括:
(5.1)在得到目标子区域的可见光图像和目标子区域的热红外图像后,云服务器基于目标子区域的热红外图像,可确定目标子区域的可见光图像中各个像素点对应的温度,由于云服务器已确定目标设备所属的设备类型,故可获取目标设备这一类设备的温度范围,接着,云服务器可在目标子区域的可见光图像中,将温度位于该温度范围外的像素点剔除,相当于在目标子区域中,将某些其余设备的一部分的可见光图像剔除,保留另一些其余设备的一部分以及目标设备的可见光图像,所保留的这一部分可见光图像可称为优化后的子区域(即优化后的检测框)的可见光图像。同理,云服务器可在目标子区域的热红外图像中,将温度位于该温度范围外的像素点剔除,相当于在目标子区域中,将某些其余设备的一部分的热红外图像剔除,保留另一些其余设备的一部分以及目标设备的热红外图像,所保留的这一部分热红外图像可称为优化后的子区域的热红外图像。
(5.2)得到优化后的子区域的可见光图像以及优化后的子区域的热红外图像后,云服务器可将优化后的子区域的可见光图像以及优化后的子区域的热红外图像输入至第四模型,以使得第四模型在优化后的子区域的可见光图像中,将另一些其余设备的一部分的可见光图像剔除,保留目标设备的可见光图像。同理,第四模型还可在优化后的子区域的热红外图像中,将另一些其余设备的一部分的热红外图像剔除,保留目标设备的热红外图像。
需要说明的是,前述的第四模型为已训练的神经网络模型,其结构如图12所示(图12为本申请实施例提供的第四模型的一个结构示意图),第四模型可包含:可见光图像编码器、热红外图像编码器、全局注意力模块、可见光图像解码器、热红外图像解码器、卷积模块(1通道卷积)。
还需要说明的是,为了训练得到第四模型,云服务器可先获取第四待训练模型(即需要训练的神经网络模型)以及一批训练数据(包含某个检测框的热红外图像以及该检测框的可见光图像),训练数据的真实处理结果(即某个设备的真实可见光图像以及真实热红外图像)。然后,云服务器可将训练数据输入至第四待训练模型,以通过第四待训练模型对训练数据进行处理,得到训练数据的预估处理结果(即该设备的预估可见光图像以及预估热红外图像)。随后,云服务器可通过预置的第四损失函数对训练数据的真实处理结果以及训练数据的预估处理结果进行计算,得到目标损失,目标损失用于指示训练数据的真实处理结果以及训练数据的预估处理结果之间的差异。最后,云服务器可基于目标损失更新第四待训练模型的参数,并利用下一批训练数据对更新参数后的第四待训练模型继续进行训练,直至满足模型训练条件(例如,目标损失收敛等等),从而得到第四模型。
应理解,在前述的步骤(1)至步骤(5)中,云服务器既可以执行所有的步骤,也可以是有选择性地执行其中某些步骤。例如,云服务器可以仅执行步骤(1)、步骤(2)、步骤(4)以及步骤(5),即云服务器先将目标区域的可见光图像对齐至目标区域的热红外图像,得到目标区域的对齐后的可见光图像。接着,云服务器对目标区域的热红外图像以及对齐后的可见光图像进行计算,得到目标区域的深度图像。然后,云服务器可基于目标区域的深度图像,对目标区域的可见光图像和目标区域的热红外图像进行分割,得到目标子区域的可见光图像和目标子区域的热红外图像,目标子区域为目标区域中,目标设备以及其余设备的一部分所占据的区域。最后,云服务器对目标子区域的可见光图像和目标子区域的热红外图像进行二次分割,得到目标设备的可见光图像和目标设备的热红外图像。
403、对目标设备的可见光图像和目标设备的热红外图像进行处理,得到温度异常区域的热红外图像。
在得到目标设备的可见光图像和目标设备的热红外图像后,云服务器可在目标设备的可见光图像和目标设备的热红外图像中,确定目标设备中温度异常区域的可见光图像和温度异常区域的热红外图像。可以理解的是,目标设备中温度异常区域的可见光图像仅呈现温度异常区域,温度异常区域的热红外图像仅呈现温度异常区域。
其中,该剔除过程包括:在得到目标设备的可见光图像和目标设备的热红外图像后,云服务器可将目标设备的可见光图像和目标设备的热红外图像输入至第五模型,以使得第五模型在目标设备的可见光图像中,将目标设备中温度正常区域的可见光图像剔除,保留目标设备中温度异常区域的可见光图像。同理,第五模型还可在目标设备的热红外图像中,将温度正常区域的热红外图像剔除,保留温度异常区域的热红外图像。
需要说明的是,前述的第五模型为已训练的神经网络模型,其结构如图13所示(图13为本申请实施例提供的第五模型的一个结构示意图),第五模型可包含:预处理模块、编码器模块、多模态特征聚合模块、全局注意力模块、解码器模块、卷积模块以及上采样模块。
还需要说明的是,为了训练得到第五模型,云服务器可先获取第五待训练模型(即需要训练的神经网络模型)以及一批训练数据(包含某个设备的热红外图像以及该设备的可见光图像),训练数据的真实处理结果(即该设备中温度异常区域的真实可见光图像以及真实热红外图像)。然后,云服务器可将训练数据输入至第五待训练模型,以通过第五待训练模型对训练数据进行处理,得到训练数据的预估处理结果(即该设备中温度异常区域的预估可见光图像以及预估热红外图像)。随后,云服务器可通过预置的第五损失函数对训练数据的真实处理结果以及训练数据的预估处理结果进行计算,得到目标损失,目标损失用于指示训练数据的真实处理结果以及训练数据的预估处理结果之间的差异。最后,云服务器可基于目标损失更新第五待训练模型的参数,并利用下一批训练数据对更新参数后的第五待训练模型继续进行训练,直至满足模型训练条件(例如,目标损失收敛等等),从而得到第五模型。
404、基于温度异常区域的热红外图像,确定温度异常区域的温度测量结果。
在得到目标设备中温度异常区域的可见光图像和温度异常区域的热红外图像后,云服务器可将温度异常区域的可见光图像和温度异常区域的热红外图像,直接作为目标设备的温度异常区域的温度测量结果,反馈给工作人员,以使得工作人员基于该温度测量结果,找到目标设备的温度异常区域,并对目标设备的温度异常区域进行检修。
进一步地,由于机器人的热成像摄像头与目标区域中的目标设备相隔一定的距离,这会导致热成像摄像头所采集到的目标设备中温度异常区域的热红外图像,无法准确指示温度异常区域的实际温度,故需要对温度异常区域的热红外图像进行温度上的修正。具体地,云服务器还设置有预设对应关系,该预设对应关系用于指示距离与温度修正值之间的对应关系。那么,云服务器在得到温度异常区域的热红外图像后,可先从目标区域的深度图像中,获取热成像摄像头与目标设备之间的距离(包含热成像摄像头与目标设备中温度异常区域之间的距离),再基于该预设对应关系以及热成像摄像头与目标设备之间的距离,确定热成像摄像头与目标设备之间的距离所对应的温度修正值,再基于该温度修正值对温度异常区域的热红外图像进行调整(例如,在温度异常区域的热红外图像上叠加该温度修正值),得到温度异常区域的调整后的热红外图像。如此一来,云服务器可将温度异常区域的可见光图像和温度异常区域的调整后的热红外图像,作为目标设备的温度异常区域的调整后的温度测量结果,并反馈给工作人员。
本申请实施例中,在工作人员的指令下,机器人可控制摄像头对包含目标设备和其余设备的目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,并发送至云服务器。接着,云服务器可对目标区域的可见光图像和目标区域的热红外图像进行处理,从而得到目标设备的可见光图像和目标设备的热红外图像。然后,云服务器可继续对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像。最后,云服务器可基于温度异常区域的热红外图像,确定目标设备的温度异常区域的温度测量结果,并上报给工作人员,以使得工作人员对目标设备的温度异常区域进行检修。前述过程过程中,云服务器在获取目标区域的可见光图像和热红外图像后,可从中将其余设备的可见光图像和热红外图像剔除,得到目标设备的可见光图像和热红外图像。由于目标设备的可见光图像和热红外图像仅呈现目标设备,在对这些图像进行分析的过程中,云服务器不会受到其余设备的影响,精准确认目标设备中温度异常区域的可见光图像和热红外图像,以此来作为目标设备的温
度异常区域的温度测量结果。由此可见,云服务器准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
进一步地,在巡检规划阶段中,云服务器可基于机器人在不同位置以及不同角度所拍摄的目标区域的图像,按照某些条件为机器人挑选出预置的位置以及预置的角度,这些条件包含:机器人位于该位置上时,摄像头与目标区域中目标设备之间的距离位于预置范围内,且机器人控制摄像头按照该角度对目标区域进行拍摄时,目标设备与其余设备的重叠程度小于预置阈值。由此可见,云服务器依照前述的条件可自动为机器人规划出最优的位置和角度,所考虑的因素较为全面,这样可以使得机器人在巡检执行阶段中,在最优的位置和角度上拍摄出最优的图像,不仅效率高,还可提高目标设备的温度测量结果的准确度。
更进一步地,在确定目标设备中温度异常区域的热红外图像后,云服务器可确定目标设备与摄像头之间的距离所对应的温度修正值,以此来调整温度异常区域的热红外图像,得到温度异常区域的调整后的热红外图像。如此一来,目标设备中温度异常区域的调整后的热红外图像更加贴近目标设备中温度异常区域的实际温度,有利于提高目标设备的温度测量结果的准确度。
更进一步地,云服务器可对目标区域的可见光图像以及热红外图像进行一系列的计算,从而自行得到目标区域的深度图像。由于目标区域的深度图像包含目标区域中各个点到摄像头之间的距离,云服务器可基于目标区域的深度图像自动获取摄像头与目标设备中温度异常区域之间的距离,以此来确定该距离对应的温度修正值,并利用温度修正值来完成温度测量结果的调整。由此可见,温度修正的整个过程可由云服务器自动完成,不需要工作人员进行操作,可降低人为操作成本。
以上是对本申请实施例提供的设备测温方法所进行的详细说明,以下将对本申请实施例提供的设备测温装置进行介绍。图14为本申请实施例提供的设备测温装置的一个结构示意图,如图14所示,该装置包括:
拍摄模块1401,用于控制摄像头对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,目标区域包含目标设备;
第一处理模块1402,用于对目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像;
第二处理模块1403,用于对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像;
确定模块1404,用于基于温度异常区域的热红外图像,确定温度异常区域的温度测量结果。
本申请实施例中,在工作人员的指令下,机器人可控制摄像头对包含目标设备和其余设备的目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,并发送至云服务器。接着,云服务器可对目标区域的可见光图像和目标区域的热红外图像进行处理,从而得到目标设备的可见光图像和目标设备的热红外图像。然后,云服务器可继续对目标设备的可见光图像和目标设备的热红外图像进行处理,得到目标设备中温度异常区域的热红外图像。最后,云服务器可基于温度异常区域的热红外图像,确定目标设备的温度异常区域的温度测量结果,并上报给工作人员,以使得工作人员对目标设备的温度异常区域进行检修。前述过程过程中,云服务器在获取目标区域的可见光图像和热红外图像后,可从中将其余设备的可见光图像和热红外图像剔除,得到目标设备的可见光图像和热红外图像。由于目标设备的可见光图像和热红外图像仅呈现目标设备,在对这些图像进行分析的过程中,云服务器不会受到其余设备的影响,精准确认目标设备中温度异常区域的可见光图像和热红外图像,以此来作为目标设备的温度异常区域的温度测量结果。由此可见,云服务器准确判定目标设备的温度异常区域,不会产生误判,从而有效提高目标设备的温度测量结果的准确度。
在一种可能实现的方式中,该装置还包括:计算模块,用于对目标区域的热红外图像以及目标区域的可见光图像进行计算,得到目标区域的深度图像;第一处理模块1402,用于对目标区域的深度图像、目标区域的可见光图像和目标区域的热红外图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该装置还包括:第三处理模块,用于将目标区域的可见光图像对齐至目标区域的热红外图像,得到目标区域的对齐后的可见光图像;计算模块,用于对目标区域的热红外图像
以及对齐后的可见光图像进行计算,得到目标区域的深度图像;第一处理模块1402,用于对目标区域的对齐后的可见光图像、目标区域的热红外图像和目标区域的深度图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该装置还包括:第四处理模块,用于基于目标区域的深度图像,在目标区域的热红外图像以及对齐后的可见光图像中,确定目标区域中前景的热红外图像以及前景的可见光图像,前景包含目标设备;第一处理模块,用于对目标区域的深度图像、前景的热红外图像以及前景的可见光图像进行处理,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,第一处理模块1402,用于:对目标区域的可见光图像和目标区域的热红外图像进行分割,得到目标子区域的可见光图像和目标子区域的热红外图像,目标子区域为目标区域中,目标设备以及其余设备的一部分所占据的区域;对目标子区域的可见光图像和目标子区域的热红外图像进行二次分割,得到目标设备的可见光图像和目标设备的热红外图像。
在一种可能实现的方式中,该装置还包括:获取模块,用于基于所述目标区域的可见光图像以及所述目标区域的热红外图像,获取所述摄像头与所述温度异常区域之间的距离;调整模块,用于基于摄像头与温度异常区域之间的距离和预设对应关系对温度测量结果进行调整,得到温度异常区域的调整后的温度测量结果,预设对应关系用于指示距离与温度修正值之间的对应关系。
在一种可能实现的方式中,拍摄模块1401,用于控制摄像头在预置的位置按照预置的角度对目标区域进行拍摄,得到目标区域的可见光图像和目标区域的热红外图像,位于位置上的摄像头与目标区域中的目标设备之间的距离位于预置范围内,摄像头按照角度对目标区域进行拍摄时,目标设备与其余设备的重叠程度小于预置阈值。
在一种可能实现的方式中,摄像头包含光成像摄像头以及热成像摄像头。
需要说明的是,上述装置各模块/单元之间的信息交互、执行过程等内容,由于与本申请方法实施例基于同一构思,其带来的技术效果与本申请方法实施例相同,具体内容可参考本申请实施例前述所示的方法实施例中的叙述,此处不再赘述。
本申请实施例还涉及一种执行设备,图15为本申请实施例提供的执行设备的一个结构示意图。如图15所示,执行设备1500具体可以表现为手机、平板、笔记本电脑、智能穿戴设备、服务器等,此处不做限定。其中,执行设备1500上可部署有图14对应实施例中所描述的设备测温装置,用于实现图4对应实施例中设备测温的功能。具体的,执行设备1500包括:接收器1501、发射器1502、处理器1503和存储器1504(其中执行设备1500中的处理器1503的数量可以一个或多个,图15中以一个处理器为例),其中,处理器1503可以包括应用处理器15031和通信处理器15032。在本申请的一些实施例中,接收器1501、发射器1502、处理器1503和存储器1504可通过总线或其它方式连接。
存储器1504可以包括只读存储器和随机存取存储器,并向处理器1503提供指令和数据。存储器1504的一部分还可以包括非易失性随机存取存储器(non-volatile random access memory,NVRAM)。存储器1504存储有处理器和操作指令、可执行模块或者数据结构,或者它们的子集,或者它们的扩展集,其中,操作指令可包括各种操作指令,用于实现各种操作。
处理器1503控制执行设备的操作。具体的应用中,执行设备的各个组件通过总线***耦合在一起,其中总线***除包括数据总线之外,还可以包括电源总线、控制总线和状态信号总线等。但是为了清楚说明起见,在图中将各种总线都称为总线***。
上述本申请实施例揭示的方法可以应用于处理器1503中,或者由处理器1503实现。处理器1503可以是一种集成电路芯片,具有信号的处理能力。在实现过程中,上述方法的各步骤可以通过处理器1503中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器1503可以是通用处理器、数字信号处理器(digital signal processing,DSP)、微处理器或微控制器,还可进一步包括专用集成电路(application specific integrated circuit,ASIC)、现场可编程门阵列(field-programmable gate array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。该处理器1503可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成,或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储
器,闪存、只读存储器,可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器1504,处理器1503读取存储器1504中的信息,结合其硬件完成上述方法的步骤。
接收器1501可用于接收输入的数字或字符信息,以及产生与执行设备的相关设置以及功能控制有关的信号输入。发射器1502可用于通过第一接口输出数字或字符信息;发射器1502还可用于通过第一接口向磁盘组发送指令,以修改磁盘组中的数据;发射器1502还可以包括显示屏等显示设备。
本申请实施例中,在一种情况下,处理器1503,用于通过图4对应实施例中的各个神经网络模型(包含第一模型、第二模型、第三模型、第四模型以及第五模型等等),完成针对目标设备的测温操作。
本申请实施例还涉及一种训练设备,图16为本申请实施例提供的训练设备的一个结构示意图。如图16所示,训练设备1600由一个或多个服务器实现,训练设备1600可因配置或性能不同而产生比较大的差异,可以包括一个或一个以***处理器(central processing units,CPU)1614(例如,一个或一个以上处理器)和存储器1632,一个或一个以上存储应用程序1642或数据1644的存储介质1630(例如一个或一个以上海量存储设备)。其中,存储器1632和存储介质1630可以是短暂存储或持久存储。存储在存储介质1630的程序可以包括一个或一个以上模块(图示没标出),每个模块可以包括对训练设备中的一系列指令操作。更进一步地,中央处理器1614可以设置为与存储介质1630通信,在训练设备1600上执行存储介质1630中的一系列指令操作。
训练设备1600还可以包括一个或一个以上电源1626,一个或一个以上有线或无线网络接口1650,一个或一个以上输入输出接口1658;或,一个或一个以上操作***1641,例如Windows ServerTM,Mac OS XTM,UnixTM,LinuxTM,FreeBSDTM等等。
具体的,训练设备可以实现图4对应实施例中的各个神经网络模型(包含第一模型、第二模型、第三模型、第四模型以及第五模型等等)的训练过程。
本申请实施例还涉及一种计算机存储介质,该计算机可读存储介质中存储有用于进行信号处理的程序,当其在计算机上运行时,使得计算机执行如前述执行设备所执行的步骤,或者,使得计算机执行如前述训练设备所执行的步骤。
本申请实施例还涉及一种计算机程序产品,该计算机程序产品存储有指令,该指令在由计算机执行时使得计算机执行如前述执行设备所执行的步骤,或者,使得计算机执行如前述训练设备所执行的步骤。
本申请实施例提供的执行设备、训练设备或终端设备具体可以为芯片,芯片包括:处理单元和通信单元,所述处理单元例如可以是处理器,所述通信单元例如可以是输入/输出接口、管脚或电路等。该处理单元可执行存储单元存储的计算机执行指令,以使执行设备内的芯片执行上述实施例描述的数据处理方法,或者,以使训练设备内的芯片执行上述实施例描述的数据处理方法。可选地,所述存储单元为所述芯片内的存储单元,如寄存器、缓存等,所述存储单元还可以是所述无线接入设备端内的位于所述芯片外部的存储单元,如只读存储器(read-only memory,ROM)或可存储静态信息和指令的其他类型的静态存储设备,随机存取存储器(random access memory,RAM)等。
具体的,请参阅图17,图17为本申请实施例提供的芯片的一个结构示意图,所述芯片可以表现为神经网络处理器NPU 1700,NPU 1700作为协处理器挂载到主CPU(Host CPU)上,由Host CPU分配任务。NPU的核心部分为运算电路1703,通过控制器1704控制运算电路1703提取存储器中的矩阵数据并进行乘法运算。
在一些实现中,运算电路1703内部包括多个处理单元(Process Engine,PE)。在一些实现中,运算电路1703是二维脉动阵列。运算电路1703还可以是一维脉动阵列或者能够执行例如乘法和加法这样的数学运算的其它电子线路。在一些实现中,运算电路1703是通用的矩阵处理器。
举例来说,假设有输入矩阵A,权重矩阵B,输出矩阵C。运算电路从权重存储器1702中取矩阵B相应的数据,并缓存在运算电路中每一个PE上。运算电路从输入存储器1701中取矩阵A数据与矩阵B进行矩阵运算,得到的矩阵的部分结果或最终结果,保存在累加器(accumulator)1708中。
统一存储器1706用于存放输入数据以及输出数据。权重数据直接通过存储单元访问控制器(Direct Memory Access Controller,DMAC)1705,DMAC被搬运到权重存储器1702中。输入数据也通过DMAC被搬运到统一存储器1706中。
BIU为Bus Interface Unit即,总线接口单元1713,用于AXI总线与DMAC和取指存储器(Instruction Fetch Buffer,IFB)1709的交互。
总线接口单元1713(Bus Interface Unit,简称BIU),用于取指存储器1709从外部存储器获取指令,还用于存储单元访问控制器1705从外部存储器获取输入矩阵A或者权重矩阵B的原数据。
DMAC主要用于将外部存储器DDR中的输入数据搬运到统一存储器1706或将权重数据搬运到权重存储器1702中或将输入数据数据搬运到输入存储器1701中。
向量计算单元1707包括多个运算处理单元,在需要的情况下,对运算电路1703的输出做进一步处理,如向量乘,向量加,指数运算,对数运算,大小比较等等。主要用于神经网络中非卷积/全连接层网络计算,如Batch Normalization(批归一化),像素级求和,对预测标签平面进行上采样等。
在一些实现中,向量计算单元1707能将经处理的输出的向量存储到统一存储器1706。例如,向量计算单元1707可以将线性函数;或,非线性函数应用到运算电路1703的输出,例如对卷积层提取的预测标签平面进行线性插值,再例如累加值的向量,用以生成激活值。在一些实现中,向量计算单元1707生成归一化的值、像素级求和的值,或二者均有。在一些实现中,处理过的输出的向量能够用作到运算电路1703的激活输入,例如用于在神经网络中的后续层中的使用。
控制器1704连接的取指存储器(instruction fetch buffer)1709,用于存储控制器1704使用的指令;
统一存储器1706,输入存储器1701,权重存储器1702以及取指存储器1709均为On-Chip存储器。外部存储器私有于该NPU硬件架构。
其中,上述任一处提到的处理器,可以是一个通用中央处理器,微处理器,ASIC,或一个或多个用于控制上述程序执行的集成电路。
另外需说明的是,以上所描述的装置实施例仅仅是示意性的,其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。另外,本申请提供的装置实施例附图中,模块之间的连接关系表示它们之间具有通信连接,具体可以实现为一条或多条通信总线或信号线。
通过以上的实施方式的描述,所属领域的技术人员可以清楚地了解到本申请可借助软件加必需的通用硬件的方式来实现,当然也可以通过专用硬件包括专用集成电路、专用CPU、专用存储器、专用元器件等来实现。一般情况下,凡由计算机程序完成的功能都可以很容易地用相应的硬件来实现,而且,用来实现同一功能的具体硬件结构也可以是多种多样的,例如模拟电路、数字电路或专用电路等。但是,对本申请而言更多情况下软件程序实现是更佳的实施方式。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品存储在可读取的存储介质中,如计算机的软盘、U盘、移动硬盘、ROM、RAM、磁碟或者光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,训练设备,或者网络设备等)执行本申请各个实施例所述的方法。
在上述实施例中,可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时,可以全部或部分地以计算机程序产品的形式实现。
所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时,全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中,或者从一个计算机可读存储介质向另一计算机可读存储介质传输,例如,所述计算机指令可以从一个网站站点、计算机、训练设备或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、训练设备或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存储的任何可用介质或者是包含一个或多个可用介质集成的训练设备、数据中心等数据存储设备。所述可用介质可以是磁性介质,(例如,软盘、硬盘、磁带)、光介质(例如,DVD)、或者半导体介质(例如固态硬盘(Solid State Disk,SSD))等。
Claims (19)
- 一种设备测温方法,其特征在于,所述方法包括:控制摄像头对目标区域进行拍摄,得到所述目标区域的可见光图像和所述目标区域的热红外图像,所述目标区域包含目标设备;对所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像;对所述目标设备的可见光图像和所述目标设备的热红外图像进行处理,得到所述目标设备中温度异常区域的热红外图像;基于所述温度异常区域的热红外图像,确定所述温度异常区域的温度测量结果。
- 根据权利要求1所述的方法,其特征在于,所述方法还包括:对所述目标区域的热红外图像以及所述目标区域的可见光图像进行计算,得到所述目标区域的深度图像;所述对所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像包括:对所述目标区域的深度图像、所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求2所述的方法,其特征在于,所述方法还包括:将所述目标区域的可见光图像对齐至所述目标区域的热红外图像,得到所述目标区域的对齐后的可见光图像;所述对所述目标区域的热红外图像以及所述目标区域的可见光图像进行计算,得到所述目标区域的深度图像包括:对所述目标区域的热红外图像以及所述对齐后的可见光图像进行计算,得到所述目标区域的深度图像;所述对所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像包括:对所述目标区域的对齐后的可见光图像、所述目标区域的热红外图像和所述目标区域的深度图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求1至3任意一项所述的方法,其特征在于,所述方法还包括:基于所述目标区域的深度图像,在所述目标区域的热红外图像以及所述对齐后的可见光图像中,确定所述目标区域中前景的热红外图像以及所述前景的可见光图像,所述前景包含目标设备;所述对所述目标区域的对齐后的可见光图像、所述目标区域的热红外图像和所述目标区域的深度图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像包括:对所述目标区域的深度图像、所述前景的热红外图像以及所述前景的可见光图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求1至4任意一项所述的方法,其特征在于,所述对所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像包括:对所述目标区域的可见光图像和所述目标区域的热红外图像进行分割,得到目标子区域的可见光图像和所述目标子区域的热红外图像,所述目标子区域为所述目标区域中,所述目标设备以及其余设备的一部分所占据的区域;对目标子区域的可见光图像和所述目标子区域的热红外图像进行二次分割,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求1至5任意一项所述的方法,其特征在于,所述方法还包括:基于所述目标区域的可见光图像以及所述目标区域的热红外图像,获取所述摄像头与所述温度异常区域之间的距离;基于所述摄像头与所述温度异常区域之间的距离和预设对应关系对所述温度测量结果进行调整,得 到所述温度异常区域的调整后的温度测量结果,所述预设对应关系用于指示所述距离与温度修正值之间的对应关系。
- 根据权利要求1至6任意一项所述的方法,其特征在于,所述控制摄像头对目标区域进行拍摄,得到所述目标区域的可见光图像和所述目标区域的热红外图像包括:控制摄像头在预置的位置按照预置的角度对目标区域进行拍摄,得到所述目标区域的可见光图像和所述目标区域的热红外图像,位于所述位置上的所述摄像头与所述目标区域中的所述目标设备之间的距离位于预置范围内,所述摄像头按照所述角度对所述目标区域进行拍摄时,所述目标设备与其余设备的重叠程度小于预置阈值。
- 根据权利要求7所述的方法,其特征在于,所述摄像头包含光成像摄像头以及热成像摄像头。
- 一种设备测温装置,其特征在于,所述装置包括:拍摄模块,用于控制摄像头对目标区域进行拍摄,得到所述目标区域的可见光图像和所述目标区域的热红外图像,所述目标区域包含目标设备;第一处理模块,用于对所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像;第二处理模块,用于对所述目标设备的可见光图像和所述目标设备的热红外图像进行处理,得到所述目标设备中温度异常区域的热红外图像;确定模块,用于基于所述温度异常区域的热红外图像,确定所述温度异常区域的温度测量结果。
- 根据权利要求9所述的装置,其特征在于,所述装置还包括:计算模块,用于对所述目标区域的热红外图像以及所述目标区域的可见光图像进行计算,得到所述目标区域的深度图像;所述第一处理模块,用于对所述目标区域的深度图像、所述目标区域的可见光图像和所述目标区域的热红外图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求10所述的装置,其特征在于,所述装置还包括:第三处理模块,用于将所述目标区域的可见光图像对齐至所述目标区域的热红外图像,得到所述目标区域的对齐后的可见光图像;所述计算模块,用于对所述目标区域的热红外图像以及所述对齐后的可见光图像进行计算,得到所述目标区域的深度图像;所述第一处理模块,用于对所述目标区域的对齐后的可见光图像、所述目标区域的热红外图像和所述目标区域的深度图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求9至11任意一项所述的装置,其特征在于,所述装置还包括:第四处理模块,用于基于所述目标区域的深度图像,在所述目标区域的热红外图像以及所述对齐后的可见光图像中,确定所述目标区域中前景的热红外图像以及所述前景的可见光图像,所述前景包含目标设备;所述第一处理模块,用于对所述目标区域的深度图像、所述前景的热红外图像以及所述前景的可见光图像进行处理,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求9至12任意一项所述的装置,其特征在于,所述第一处理模块,用于:对所述目标区域的可见光图像和所述目标区域的热红外图像进行分割,得到目标子区域的可见光图像和所述目标子区域的热红外图像,所述目标子区域为所述目标区域中,所述目标设备以及其余设备的一部分所占据的区域;对目标子区域的可见光图像和所述目标子区域的热红外图像进行二次分割,得到所述目标设备的可见光图像和所述目标设备的热红外图像。
- 根据权利要求9至13任意一项所述的装置,其特征在于,所述装置还包括:获取模块,用于基于所述目标区域的可见光图像以及所述目标区域的热红外图像,获取所述摄像头与所述温度异常区域之间的距离;调整模块,用于基于所述摄像头与所述温度异常区域之间的距离和预设对应关系对所述温度测量结果进行调整,得到所述温度异常区域的调整后的温度测量结果,所述预设对应关系用于指示所述距离与 温度修正值之间的对应关系。
- 根据权利要求9至14任意一项所述的装置,其特征在于,所述拍摄模块,用于控制摄像头在预置的位置按照预置的角度对目标区域进行拍摄,得到所述目标区域的可见光图像和所述目标区域的热红外图像,位于所述位置上的所述摄像头与所述目标区域中的所述目标设备之间的距离位于预置范围内,所述摄像头按照所述角度对所述目标区域进行拍摄时,所述目标设备与其余设备的重叠程度小于预置阈值。
- 根据权利要求9至15任意一项所述的装置,其特征在于,所述摄像头包含光成像摄像头以及热成像摄像头。
- 一种设备测温装置,其特征在于,所述装置包括存储器和处理器;所述存储器存储有代码,所述处理器被配置为执行所述代码,当所述代码被执行时,所述设备测温装置执行如权利要求1至8任意一项所述的方法。
- 一种计算机存储介质,其特征在于,所述计算机存储介质存储有一个或多个指令,所述指令在由一个或多个计算机执行时使得所述一个或多个计算机实施权利要求1至8任一所述的方法。
- 一种计算机程序产品,其特征在于,所述计算机程序产品存储有指令,所述指令在由计算机执行时,使得所述计算机实施权利要求1至8任意一项所述的方法。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211427331.4A CN118050087A (zh) | 2022-11-15 | 2022-11-15 | 一种设备测温方法及其相关设备 |
CN202211427331.4 | 2022-11-15 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2024104365A1 true WO2024104365A1 (zh) | 2024-05-23 |
Family
ID=91045554
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2023/131686 WO2024104365A1 (zh) | 2022-11-15 | 2023-11-15 | 一种设备测温方法及其相关设备 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN118050087A (zh) |
WO (1) | WO2024104365A1 (zh) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109029731A (zh) * | 2018-05-24 | 2018-12-18 | 河海大学常州校区 | 一种基于多目视觉的电力设备异常监测***及方法 |
CN110987189A (zh) * | 2019-11-21 | 2020-04-10 | 北京都是科技有限公司 | 对目标对象进行温度检测的方法、***以及装置 |
CN112257664A (zh) * | 2020-11-12 | 2021-01-22 | Oppo广东移动通信有限公司 | 图像融合方法、装置、计算机设备和存储介质 |
WO2021232587A1 (zh) * | 2020-05-21 | 2021-11-25 | 平安国际智慧城市科技股份有限公司 | 基于图像处理的双光摄像机测温方法及相关设备 |
CN114485953A (zh) * | 2020-11-13 | 2022-05-13 | 杭州海康威视数字技术股份有限公司 | 温度测量方法、装置及*** |
-
2022
- 2022-11-15 CN CN202211427331.4A patent/CN118050087A/zh active Pending
-
2023
- 2023-11-15 WO PCT/CN2023/131686 patent/WO2024104365A1/zh unknown
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109029731A (zh) * | 2018-05-24 | 2018-12-18 | 河海大学常州校区 | 一种基于多目视觉的电力设备异常监测***及方法 |
CN110987189A (zh) * | 2019-11-21 | 2020-04-10 | 北京都是科技有限公司 | 对目标对象进行温度检测的方法、***以及装置 |
WO2021232587A1 (zh) * | 2020-05-21 | 2021-11-25 | 平安国际智慧城市科技股份有限公司 | 基于图像处理的双光摄像机测温方法及相关设备 |
CN112257664A (zh) * | 2020-11-12 | 2021-01-22 | Oppo广东移动通信有限公司 | 图像融合方法、装置、计算机设备和存储介质 |
CN114485953A (zh) * | 2020-11-13 | 2022-05-13 | 杭州海康威视数字技术股份有限公司 | 温度测量方法、装置及*** |
Also Published As
Publication number | Publication date |
---|---|
CN118050087A (zh) | 2024-05-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2019223382A1 (zh) | 单目深度估计方法及其装置、设备和存储介质 | |
WO2022042713A1 (zh) | 一种用于计算设备的深度学习训练方法和装置 | |
WO2022179581A1 (zh) | 一种图像处理方法及相关设备 | |
JP6798183B2 (ja) | 画像解析装置、画像解析方法およびプログラム | |
US20220375213A1 (en) | Processing Apparatus and Method and Storage Medium | |
US12039440B2 (en) | Image classification method and apparatus, and image classification model training method and apparatus | |
CN111368972B (zh) | 一种卷积层量化方法及其装置 | |
CN110222718B (zh) | 图像处理的方法及装置 | |
WO2021103731A1 (zh) | 一种语义分割方法、模型训练方法及装置 | |
WO2023083030A1 (zh) | 一种姿态识别方法及其相关设备 | |
CN112598597A (zh) | 一种降噪模型的训练方法及相关装置 | |
WO2022179606A1 (zh) | 一种图像处理方法及相关装置 | |
CN112258565B (zh) | 图像处理方法以及装置 | |
WO2022052782A1 (zh) | 图像的处理方法及相关设备 | |
WO2022111387A1 (zh) | 一种数据处理方法及相关装置 | |
CN111950700A (zh) | 一种神经网络的优化方法及相关设备 | |
US20240185568A1 (en) | Image Classification Method and Related Device Thereof | |
WO2022165722A1 (zh) | 单目深度估计方法、装置及设备 | |
CN114359289A (zh) | 一种图像处理方法及相关装置 | |
CN111738403A (zh) | 一种神经网络的优化方法及相关设备 | |
CN114677422A (zh) | 深度信息生成方法、图像虚化方法和视频虚化方法 | |
CN115239581A (zh) | 一种图像处理方法及相关装置 | |
WO2024140973A1 (zh) | 一种动作计数方法及其相关设备 | |
CN112528978B (zh) | 人脸关键点的检测方法、装置、电子设备及存储介质 | |
WO2024046144A1 (zh) | 一种视频处理方法及其相关设备 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23890798 Country of ref document: EP Kind code of ref document: A1 |