WO2022105231A1 - 入库管理方法、装置、仓库管理***和电子*** - Google Patents

入库管理方法、装置、仓库管理***和电子*** Download PDF

Info

Publication number
WO2022105231A1
WO2022105231A1 PCT/CN2021/102425 CN2021102425W WO2022105231A1 WO 2022105231 A1 WO2022105231 A1 WO 2022105231A1 CN 2021102425 W CN2021102425 W CN 2021102425W WO 2022105231 A1 WO2022105231 A1 WO 2022105231A1
Authority
WO
WIPO (PCT)
Prior art keywords
target stack
target
stack
goods
conveying
Prior art date
Application number
PCT/CN2021/102425
Other languages
English (en)
French (fr)
Inventor
陈德平
孙伟
杨磊
王银学
童孝康
Original Assignee
北京旷视机器人技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京旷视机器人技术有限公司 filed Critical 北京旷视机器人技术有限公司
Publication of WO2022105231A1 publication Critical patent/WO2022105231A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders

Definitions

  • the present disclosure relates to the technical field of stacking and warehousing, and in particular, to a warehousing management method, device, warehouse management system and electronic system.
  • the traditional stacking warehousing mainly relies on the photo grating to detect the stacking type of the stacking, and the quantity of the goods is checked and checked and the shape of the stacking is mainly detected by manual observation.
  • the general method is to form a protective light curtain by emitting infrared rays from a safety light curtain composed of a light grating, to perform two-dimensional imaging of the goods, to analyze the imaging results, and finally to determine whether the stacking type of the stack is qualified.
  • this method requires high labor and time costs, and the detection efficiency and detection accuracy are both low.
  • the purpose of the present disclosure is to provide a warehouse management method, device, warehouse management system and electronic system, so as to save manpower and time resources, and improve detection efficiency and detection accuracy.
  • an embodiment of the present disclosure provides a warehouse management method, the method is applied to a server, and the server is communicatively connected to a plurality of camera devices; the method includes: in the process of the target stack being transferred on the transfer device , control multiple camera devices to collect multi-angle images of the target stack; among them, multiple camera devices are set on the conveying path of the target stack to shoot the target stack from multiple different angles; multi-angle images of the target stack Perform image analysis, and determine whether the target stack is normally put into storage according to the analysis results.
  • the above-mentioned conveying device is provided with a photoelectric sensor, and the photoelectric sensor is configured to detect whether the target stack reaches a designated monitoring point; the above-mentioned controlling multiple camera devices to collect multi-angle images of the target stack includes: when the photoelectric sensor detects When the target stack reaches the designated monitoring point, an image acquisition instruction is sent to multiple camera devices, so that the multiple camera devices collect multi-angle images of the target stack.
  • the above-mentioned conveying equipment includes a linear conveying line, and a plurality of imaging devices are arranged on the gantry corresponding to the linear conveying line; the above-mentioned designated monitoring point is a monitoring point corresponding to the gantry.
  • the number of multiple camera devices is five, wherein four camera devices are respectively configured to photograph the four sides of the target stack located at the monitoring point, and one camera device is configured to photograph the top surface of the target stack.
  • the above-mentioned conveying device includes a first conveying line and a second conveying line constituting an L-shaped conveying line
  • the above-mentioned photoelectric sensor includes a first photoelectric sensor and a second photoelectric sensor
  • the first photoelectric sensor and the second photoelectric sensor are respectively arranged at On the first conveying line and the second conveying line, the first group of imaging devices among the plurality of imaging devices is arranged on the first gantry of the first conveying line, and the second group of imaging devices among the plurality of imaging devices is arranged on the second gantry.
  • sending an image acquisition instruction to multiple camera devices includes: when the first photoelectric sensor detects that the target stack arrives at the first conveying point When the designated monitoring point on the line, send an image acquisition instruction to the first group of camera equipment; when the second photoelectric sensor detects that the target stack reaches the designated monitoring point on the second conveyor line, send image acquisition to the second group of camera equipment. instruction.
  • the number of the plurality of imaging devices is 5, wherein, the first group of imaging devices includes 3 imaging devices, which are respectively configured to photograph the two sides and The top surface of the target stack; the second group of imaging devices includes two imaging devices, which are respectively configured to photograph the other two side surfaces of the target stack located at the designated monitoring point on the second conveying line.
  • performing the image analysis on the multi-angle images of the target stacking includes: performing at least one of the following detections on the multi-angle images of the target stacking through a pre-trained image analysis model: detection of the quantity of goods, detection of the stacking type. , Deformation detection of goods packaging, damage detection of goods packaging and detection of stains on goods packaging; determine whether the target stacking is abnormal according to the test results.
  • the above-mentioned image analysis includes stacking type detection
  • the above-mentioned stacking type detection includes at least one of the following: detecting the inclination of the target stack; detecting the placement relationship between the goods in the target stack and the pallet; The relationship between the top and bottom of the goods; the distance between the goods in the target stack is detected.
  • the method before performing the image analysis on the multi-angle images of the target stack, the method further includes: acquiring the initial quantity of the target stack to be put into storage;
  • the above-mentioned image analysis of the multi-angle images of the target stack includes: determining the actual number of goods contained in the target stack according to the multi-angle images of the target stack; judging whether the actual number of goods is consistent with the initial number of goods; if not, determining the target Stacking exception.
  • the multi-angle image includes side images of 4 sides and a top image of one top surface of the target stack; the above step of determining the actual number of goods included in the target stack according to the multi-angle image of the target stack, Including: respectively taking each side image and top image as a target image, and marking in each target image the product frame contained on the stacking surface closest to the target camera device; wherein, the target camera device is the camera device that collects the target image ; Determine the actual number of items contained in the target stack based on the item boxes contained in each target image after labeling.
  • the above-mentioned determining whether the target stack is normally put into storage according to the analysis result includes: if the analysis result is that the target stack is abnormal, controlling the conveying equipment to transfer the target stack to the abnormal processing area; if the analysis result is that the target stack is normal , control the conveying equipment to put the target stack into the warehouse normally.
  • an embodiment of the present disclosure also provides a storage management device, the device is applied to a server, and the server is communicatively connected to a plurality of camera devices; the device includes: a first control module, configured to be configured to be in the delivery process of the target stack. In the process of conveying on the equipment, control multiple camera devices to collect multi-angle images of the target stack; multiple camera devices are set on the conveying path of the target stack to shoot the target stack from multiple different angles; the second control The module is configured to perform image analysis on the multi-angle images of the target stack, and determine whether the target stack is normally put into storage according to the analysis result.
  • the second control module is further configured to: perform at least one of the following detections on the multi-angle images of the target stack by using a pre-trained image analysis model: detection of the quantity of goods, detection of stacking types, and packaging of goods. Deformation detection, product packaging damage detection, and product packaging stain detection; determine whether the target stack is abnormal according to the detection results.
  • the device further includes: an information acquisition module configured to: acquire the initial quantity of the target stack to be put into storage; the second control module is further configured to: according to the multi-angle image of the target stack Determine the actual number of goods contained in the target stack; determine whether the actual number of goods is consistent with the initial number of goods; if inconsistent, determine that the target stack is abnormal.
  • an information acquisition module configured to: acquire the initial quantity of the target stack to be put into storage
  • the second control module is further configured to: according to the multi-angle image of the target stack Determine the actual number of goods contained in the target stack; determine whether the actual number of goods is consistent with the initial number of goods; if inconsistent, determine that the target stack is abnormal.
  • the second control module is further configured to: if the analysis result is that the target stack is abnormal, control the conveying device to transfer the target stack to the abnormal processing area; if the analysis result is that the target stack is abnormal The stack is normal, and the conveying device is controlled to put the target stack into the warehouse normally.
  • an embodiment of the present disclosure further provides a warehouse management system, including a server, a conveying device communicatively connected to the server, and a plurality of camera devices; the conveying device is configured to convey target stacks under the control of the server; the server is configured to: During the process of the target stack being conveyed on the conveying equipment, control multiple camera devices to collect multi-angle images of the target stack; and perform image analysis on the multi-angle images of the target stack, and determine whether to stack the target according to the analysis results. Normal warehousing; multiple camera devices are set on the conveying path of the target stack to shoot the target stack from multiple different angles.
  • the above-mentioned conveying device includes a linear conveying line, and a plurality of imaging devices are arranged on the gantry corresponding to the linear conveying line; or, the conveying device includes an L-shaped conveying line, and one of the multiple imaging devices The first group of imaging devices is arranged on the first gantry corresponding to the L-shaped conveying line, and the second group of imaging devices among the plurality of imaging devices is arranged on the second gantry corresponding to the L-shaped conveying line.
  • an embodiment of the present disclosure further provides an electronic system, the electronic system includes: a processing device and a storage device; the storage device stores a computer program, and the computer program executes the above storage management method when the processed device runs.
  • embodiments of the present disclosure further provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is run by a processing device, the steps of the above-mentioned storage management method are executed.
  • the multi-angle images of the target stack are collected by multiple camera devices during the transfer process of the target stack. Perform image analysis on multi-angle images, and determine whether the target stack is normally put into storage according to the analysis results.
  • the server analyzes the multi-angle images of the target stack to determine whether it is normally put into storage, and then does not need to manually perform stacking inspection, which can save labor costs and time costs; multi-angle images are detected by image analysis. , Accurately determine whether the target stack can be normally put into storage, thereby improving the detection efficiency and detection accuracy of stack detection.
  • FIG. 1 is a schematic structural diagram of an electronic system according to an embodiment of the present disclosure
  • FIG. 2 is a flowchart of a storage management method provided by an embodiment of the present disclosure
  • FIG. 3 is a schematic diagram of an arrangement of a camera device according to an embodiment of the present disclosure.
  • FIG. 4 is a schematic diagram of another setting manner of a camera device according to an embodiment of the present disclosure.
  • FIG. 5 is a flowchart of another storage management method provided by an embodiment of the present disclosure.
  • FIG. 6 is a schematic diagram of a target stack with a risk of sloping scattering according to an embodiment of the present disclosure
  • FIG. 7 is a schematic diagram of a stacking abnormality of a target stacking according to an embodiment of the present disclosure.
  • FIG. 8 is a flowchart of another storage management method provided by an embodiment of the present disclosure.
  • FIG. 9 is a schematic diagram of a target image provided by an embodiment of the present disclosure.
  • FIG. 10 is a schematic overall flowchart of a storage management method provided by an embodiment of the present disclosure.
  • FIG. 11 is a schematic structural diagram of a storage management apparatus according to an embodiment of the present disclosure.
  • the quantity and shape detection of goods are mainly carried out manually when stacking and warehousing, and the efficiency and accuracy of manual counting and shape detection are not high.
  • manual counting requires high labor and time costs, low efficiency, and poor continuity, and the manual counting is not sufficient for large warehouse detection when there are too many goods; manual detection of the accuracy of the stacking shape It is difficult to find all the smudges on the stacking surface, and it is impossible to ensure that the quality of the inbound stacking meets the requirements; in addition, the two-dimensional imaging of the goods through the light curtain has great limitations, requiring high cost and high cost. Only the density grating equipment can obtain higher detection accuracy; the unqualified stacks will cost a lot of money to be taken out after being put into the warehouse. If the processing is not timely, additional losses may be caused, and the time cost and labor cost of processing unqualified stacks will also be high. very high.
  • the embodiments of the present disclosure provide a warehouse management method, device, warehouse management system and electronic system.
  • the technology can be applied to electronic systems such as mobile phones, computers, and servers that need to perform stacking detection. .
  • the electronic system can be used to implement the warehouse management method and device and the warehouse management system according to the embodiments of the present disclosure.
  • FIG. 1 is a schematic structural diagram of an electronic system
  • the electronic system 100 includes one or more processing devices 102 , one or more storage devices 104 , an input device 106 , an output device 108 and one or more image acquisition devices 110 , these components are interconnected by a bus system 112 and/or other forms of connection mechanisms (not shown).
  • bus system 112 and/or other forms of connection mechanisms (not shown).
  • the components and structures of the electronic system 100 shown in FIG. 1 are only exemplary and not limiting, and the electronic system may also have other components and structures as required.
  • the processing device 102 can be a server, an intelligent terminal, or a device that includes a central processing unit (CPU) or other forms of processing units with data processing capabilities and/or instruction execution capabilities, and can process data from other components in the electronic system 100 . Processing, other components in the electronic system 100 may also be controlled to perform inventory management functions.
  • CPU central processing unit
  • other components in the electronic system 100 may also be controlled to perform inventory management functions.
  • Storage 104 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and/or non-volatile memory.
  • Volatile memory may include, for example, random access memory (RAM) and/or cache memory, among others.
  • Non-volatile memory may include, for example, read only memory (ROM), hard disk, flash memory, and the like.
  • One or more computer program instructions may be stored on the computer-readable storage medium, and the processing device 102 may execute the program instructions to implement the client functions (implemented by the processing device) in the following embodiments of the present disclosure and/or other desired Function.
  • client functions implemented by the processing device
  • Various application programs and various data such as various data used and/or generated by the application program, etc., may also be stored in the computer-readable storage medium.
  • Input device 106 may be a device used by a user to input instructions, and may include one or more of a keyboard, mouse, microphone, touch screen, and the like.
  • the output device 108 may output various information (eg, images or sounds) to the outside (eg, a user), and may include one or more of a display, a speaker, and the like.
  • the image acquisition device 110 (also referred to as a camera device, such as a camera, etc.) can acquire multi-angle images of the target stack, and store the acquired images in the storage device 104 for use by other components.
  • each device for implementing the warehouse management method, device, warehouse management system and electronic system may be set in an integrated manner, or may be set in a decentralized manner, such as the processing device 102, the storage device 104, the input The device 106 and the output device 108 are integrated into one body, and the image capturing device 110 is arranged at a designated position where images can be captured.
  • the electronic system can be implemented as a smart terminal such as a camera, a smart phone, a tablet computer, a computer, a vehicle-mounted terminal, and the like.
  • This embodiment provides a warehouse management method, which is applied to a server, that is, the method can be executed by a server, and the server can be a control device in a warehouse management system, which is connected with a transmission device, a camera in the warehouse management system Devices and other communication connections
  • the imaging device in the embodiment of the present disclosure may be a device that can perform image acquisition, such as a camera, a monitor, etc., and the imaging device is usually set at a designated monitoring point of the stacking conveying path.
  • the storage management method mainly includes the following steps S202 to S204:
  • Step S202 during the process of the target stack being conveyed on the conveying device, control multiple camera devices to collect multi-angle images of the target stack.
  • a plurality of camera devices are arranged on the conveying path of the target stack, and respectively photograph the target stack from a plurality of different angles.
  • the goods When there are goods that need to be put into storage, the goods are usually placed on pallets to form a stack, and then the stack is placed on a conveying device to be transported into storage.
  • the stacks placed on the conveying equipment that need to be put into storage are called target stacks.
  • the conveying equipment may be a conveying belt or a conveying trolley or the like. During the conveying process of the target stack, the target stack will be transported along the stack conveying path. Since there are multiple camera devices set at the designated monitoring points of the stack conveying path, in practical applications, one designated monitoring point can be used. Multiple camera devices with different orientations may be set, or at least one camera device may be set at multiple designated monitoring points respectively, which is not limited here. By arranging multiple camera devices, multi-angle images of the target stack can be captured while the target stack is being conveyed.
  • Multi-angle images include multiple images obtained by shooting the target stack from multiple different angles; images corresponding to multiple visible surfaces of the target stack can be captured through different angles, and the visible surface is visible to the human eye during the conveying process.
  • s surface For example: Suppose the target stack is a cube, there are 6 surfaces in total, except the lower surface which is invisible to the human eye during the conveying process, the remaining 5 surfaces (upper surface, front surface, rear surface, left surface and right surface) are visible noodle.
  • multiple camera devices may collect multi-angle images of target stacks at different positions at different times, and the images collected by each camera device may correspond to the same or different target stack surfaces. In order to ensure the accuracy of subsequent detection, images corresponding to all visible surfaces of the target stack can be collected.
  • the target stack is 2 meters away from the stacking conveying path in the 3rd second, and the camera device 1 can capture the image of the upper surface of the target stack; the target stack is 15 meters away from the stacking conveying path in the 10th second, and the camera The device 2 can capture the image of the right surface of the target stack; the target stack is 30 meters away from the stacking conveying path in the 20th second, and the camera device 3 can also capture the image of the upper surface of the target stack.
  • the server can control the camera device to enable the acquisition function when the current position of the target stack is relatively close to the location of the camera device, and then collect multiple different information about the target stack. Angled image.
  • the current position of the target stack may be determined according to the conveying speed of the conveying device, or may be determined according to the detection information fed back by the detection sensor installed on the conveying device, etc., which is not limited to the embodiment of the present disclosure.
  • step S204 image analysis is performed on the multi-angle images of the target stack, and whether the target stack is normally put into storage is determined according to the analysis result.
  • the server can perform image analysis on the multi-angle images, and the image analysis can include whether the goods contained in the target stack are properly placed, whether the number of goods is correct, and whether the outer packaging of the goods is damaged or stained, etc.
  • the means of image analysis may include image detection, image recognition, image segmentation, etc.
  • the length, width and height of the stack can be calculated according to the size of the pallet, and the volume of the stack can be calculated, and then further Calculate the number of items stacked in the stack. It is also possible to input multi-angle images into the pre-trained quantity detection model, and the quantity detection model outputs the quantity of the target stack.
  • the above method collects multi-angle images of the target stack by a plurality of camera devices during the conveying process of the target stack, performs image analysis on the multi-angle images of the target stack, and determines whether the target stack is normally put into storage according to the analysis result.
  • the server analyzes the multi-angle images of the target stack to determine whether it is normally put into storage, and then does not need to manually perform stacking inspection, which can save labor costs and time costs; multi-angle images are detected by image analysis. , Accurately determine whether the target stack can be normally put into storage, thereby improving the detection efficiency and detection accuracy of stack detection.
  • the conveying equipment in this embodiment is provided with a photoelectric sensor, and the photoelectric sensor is configured to detect whether the target stack reaches a designated monitoring point; accordingly, the above-mentioned control of multiple camera equipment to collect the target
  • the multi-angle image of the stack includes: when the above-mentioned photoelectric sensor detects that the target stack reaches the designated monitoring point, sending an image acquisition instruction to multiple camera devices, so that the multiple camera devices capture the multi-angle image of the target stack.
  • the server in this embodiment may receive prompt information sent by the photoelectric sensor when monitoring that the target stack arrives at a designated monitoring point, and send image acquisition instructions to multiple camera devices according to the prompt information. In this way, the photoelectric sensor on the conveying device detects whether the stack has reached the designated monitoring point, and then triggers the camera device for image acquisition, which can more accurately control the timing of triggering the camera device, thereby ensuring the quality of the collected images.
  • the above-mentioned multiple camera devices may be arranged on the gantry corresponding to the linear conveying line; the above-mentioned designated monitoring point is the monitoring point corresponding to the gantry.
  • multiple camera equipment can be installed on a gantry corresponding to the linear conveyor line.
  • the number of multiple camera equipment is 5, of which 4 camera equipments are configured respectively The four sides of the target stack located at the monitoring point are photographed, and one camera device is used to photograph the top surface of the target stack.
  • the distance between each of the four camera devices and the side to be photographed is within a first preset distance interval (for example: 0.9m to 1.2m, optionally, the distance is set to 1m ), the distance from the top surface of the target stack is within the second preset distance interval (for example: 0.1 m to 0.3 m, optional, the distance is set to 0.2 m), and the shooting angle corresponding to the horizontal plane is within the first preset distance range.
  • Set the angle range (for example: 40 degrees to 50 degrees, optional, the angle is set to 45 degrees).
  • Another camera device for photographing the top surface of the target stack has a distance from the top surface of the target stack within the second preset distance range, and its shooting angle is directed vertically to the target stack.
  • the conveying device includes a first conveying line and a second conveying line constituting the L-shaped conveying line
  • the above-mentioned photoelectric sensor includes a first photoelectric sensor and a second photoelectric sensor, wherein the first photoelectric sensor and the second photoelectric sensor are respectively arranged on the first conveying line and the second conveying line, the first group of imaging devices in the plurality of imaging devices is arranged on the first gantry of the first conveying line, the first group of imaging devices in the plurality of imaging devices is arranged on the first gantry of the first conveying line, Two groups of camera devices are arranged on the second gantry of the second conveying line; correspondingly, when the photoelectric sensor detects that the target stack reaches the designated monitoring point, the image acquisition instructions are sent to multiple camera devices, including: when the first When the photoelectric sensor detects that the target stack reaches the designated monitoring point on the first conveying line, it sends an image acquisition instruction to the first group of camera equipment;
  • the first photoelectric sensor when it detects the target stack, it sends a prompt message to the server to remind the target stack that the target stack has been transported to the first monitoring point corresponding to the first gantry; When the target is stacked, a prompt message is sent to the server to indicate that the target stack has been transported to the second monitoring point corresponding to the second gantry.
  • the prompt information may carry the identification information of the photoelectric sensor, or carry the pre-specified information identifier, for the purpose of prompting the current location of the target stack.
  • the prompt information received by the server is that the target stack has been transported to the first monitoring point corresponding to the first gantry, it sends image acquisition instructions to the first group of camera devices respectively;
  • the second monitoring point corresponds to the second gantry, the image acquisition instructions are respectively sent to the second group of camera devices.
  • the number of the above-mentioned multiple imaging devices may also be 5, wherein the first group of imaging devices includes 3 imaging devices, which are respectively used to photograph designated monitoring points located on the first conveying line.
  • the camera equipment that shoots two sides the distance from the side to be photographed is within the third preset distance interval, and the distance from the top surface of the target stack is within the fourth preset distance interval, The shooting angle corresponding to the horizontal plane is within the second preset angle interval; the distance between the camera equipment that shoots the top surface of the target stack in the first group of camera equipment and the top surface is within the fourth preset distance interval, and the shooting angle points vertically to the target Stacking;
  • the camera equipment in the second group of camera equipment the distance from the side to be photographed is within the third preset distance interval, the distance from the top surface of the target stack is within the fourth preset distance interval, and the distance corresponding to the horizontal plane is within the third preset distance interval.
  • the shooting angle is within the second preset angle range.
  • the first preset distance interval, the second preset distance interval, the third preset distance interval and the fourth preset distance interval may be the same or different from each other, and the first preset angle interval and the second preset angle interval It can be the same or different, and can be set flexibly according to the actual situation, which is not limited here.
  • the camera device 1 to the camera device 5 can be cameras.
  • a designated gantry of the warehouse can be set on the stacking conveying road, and the camera device 1 to the camera device 5 is set on the designated gantry.
  • the camera equipment 1 to camera equipment 4 are respectively 1 meter away from the edge of the pallet, the height can be 0.2 meters higher than the maximum passing height of the target stack, and the angle can be 45 degrees from the horizontal plane. Angle down.
  • the camera device 5 can be set just above the center of the pallet and the stack, and the height is 1 meter higher than the maximum passing height of the target stack, and shoot vertically downward.
  • camera equipment 1 , camera equipment 2 , and camera equipment 5 are arranged on the first gantry, and the images of the two sides and the top surface of the target stack are collected respectively.
  • the camera devices 1 and 2 are 1 meter away from the pallet and the stack, and the height is half of the maximum passing height of the target stack, and the angle is kept horizontal. It is 1 meter higher than the maximum passing height of the target stack, and the angle is vertically downward.
  • the pallet continues to run with the "L-shaped" conveyor belt after passing through the first gantry, the absolute direction of the stack relative to the ground remains unchanged, the running direction of the conveyor belt rotates 90 degrees clockwise, and the camera equipment 3 and camera equipment 4 are set on the second gantry. , respectively collect multi-angle images of the other two sides of the target stack.
  • the imaging device 1 may also be arranged on the second gantry, which is not limited here. In the above two setting methods, the multi-angle images of the target stack can be collected by setting fewer gantry, and in addition, the camera equipment can also be set on a larger number of gantry.
  • this embodiment provides another storage management method, which is applied to the above-mentioned server, and focuses on the image analysis method based on the image analysis model.
  • the flow chart of the method mainly includes the following steps S502 to S510:
  • Step S502 in the process of the target stack being conveyed on the conveying device, control multiple camera devices to collect multi-angle images of the target stack; The target stacking is photographed from an angle; in step S502, reference may be made to the relevant content of step S202, which will not be repeated here.
  • Step S504 perform at least one of the following detections on the multi-angle images of the target stack through the pre-trained image analysis model: commodity quantity detection, stacking type detection, commodity packaging deformation detection, commodity packaging damage detection, and commodity packaging smudge detection .
  • the stack has a cubic structure with a total of 6 surfaces.
  • the target stack is transported by pallets on the conveying equipment. Since the lower surface of the target stack is close to the pallet, it is impossible to capture images of the lower surface. Therefore, in order to comprehensively detect whether the target stack is abnormal, images corresponding to 5 surfaces can be collected by the camera device, specifically, the side images of the 4 sides and the top image of the top surface of the target stack. Image analysis is performed on these five images through a pre-trained image analysis model.
  • the image analysis model can be trained based on the image samples corresponding to each stack, and the specific training process can refer to the method of related model training, which will not be described in detail here.
  • the quantity of items in each stack can be a predetermined quantity, which is usually fixed. Therefore, the above-mentioned item quantity detection can be based on the multi-angle image of the target stack. The actual number of goods. If the actual number of goods is consistent with the specified quantity, the number of goods in the target stack is normal.
  • the stack shape in the embodiment of the present disclosure is the shape of the stack
  • the server may determine the stack shape of the target stack according to the multi-angle image of the target stack.
  • the stacking type it can be seen whether the target stacking has problems such as looseness and dislocation, so as to determine whether the target stacking has the risk of tilting and scattering.
  • FIG. 6 the schematic diagram of a target stack with a risk of sloping and falling as shown in FIG. 6, it can be seen that the upper target stacks in FIG. 6 are neatly arranged, and there is no risk of slanting and scattering; while the lower target stacks are scattered , there is a risk of tilting and scattering.
  • the above-mentioned stacking type detection may include at least one of the following:
  • the inclination of the target stack can be determined by the side image of the stack. If the side edge of the stack in the side image is no longer a straight line, but consists of multiple small line segments, it means that the stack is inclined, and the adjacent small line segments are between The greater the distance, the greater the degree of inclination, and the greater the risk of the stacking being prone to incline and scatter at this time, which is regarded as abnormal stacking.
  • the stack shape shown by 3 in FIG. 7 has a larger degree of inclination.
  • the stacking type shown in FIG. 7 that does not meet the requirements of the stacking type also includes 2) a schematic diagram of irregular stacking and 5) a schematic diagram of the top heavy and the bottom light.
  • the stacking type shown in FIG. 7 that does not meet the requirements of the stacking type also includes 2) a schematic diagram of irregular stacking and 5) a schematic diagram of the top heavy and the bottom light.
  • it can also be inferred by detecting the type of goods in the upper part of the stack and the type of goods in the lower part of the stack. Different types of goods usually have different weights of a single box. If the heavier goods are located in the upper part of the stack, it is easy to cause backlog deformation of the lower part of the goods and damage to the packaging, and it is also easy to fall during the transfer process.
  • the target stack does not meet at least one preset stacking type requirement, it means that the target stack has a stacking problem, so it can be identified as an abnormal stack.
  • the above-mentioned method provided by the embodiments of the present disclosure can determine the stacking type, quantity, and whether the packaging of the target stacking is damaged from the multi-angle images of the target stacking, and can detect abnormal stacking.
  • the above several stacking detection methods can be used at the same time or not at the same time.
  • the above several stacking detection methods can be used at the same time or not at the same time.
  • there is one detection method that identifies the target stacking as abnormal stacking even if other detection methods identify the target stacking
  • the stacking is normal stacking, and the target stacking is also abnormal stacking.
  • Step S506 determine whether the target stack is abnormal. If it is normal, go to step S508; if it is abnormal, go to step S510.
  • step S508 the conveying equipment is controlled to normally put the target stack into the warehouse.
  • Step S510 controlling the conveying device to convey the target stack to the exception handling area.
  • the exception handling area is the placement area of the abnormal target stack. After the target stack is transferred to the exception handling area, the staff can check the target stack in the exception handling area.
  • the server can send an early warning message to the warehouse staff, indicating that the target stacking is abnormal or the target stacking is stored in the abnormal processing area.
  • the early warning information carries the identification information of the target stack, and can also carry the image of the target stack (such as one or more of the above-mentioned multi-angle images). If the number of goods in the target stack is abnormal, it can also carry the target stack. the actual quantity of the item.
  • the above identification information may be a location identification, a serial number identification, or an abnormal cause identification, etc., and the abnormal target stack may be identified. After receiving the warning information, the staff can go to the exception handling area to view the target stack.
  • normal target stacking For normal target stacking, it can be stored normally through the conveying equipment, so that normal target stacking and abnormal target stacking can be placed separately, without the need for staff to manually separate abnormal target stacking and normal target stacking. .
  • the multi-angle images of the target stack are collected by a plurality of camera devices, and the multi-angle images of the target stack are input into the image analysis model to perform the detection of the quantity of goods, the detection of the shape of the stack, the detection of the deformation of the packaging of the goods, the detection of the damage of the packaging of the goods, or the packaging of the goods.
  • Smudge detection if the target stack is abnormal, the server controls the conveying device to transfer the target stack to the abnormal processing area.
  • This method does not require manual stacking detection, which can save labor costs and time costs; detecting multi-angle images through image analysis can quickly and accurately determine whether the target stacking is abnormal, thereby improving the detection efficiency and detection accuracy of stacking detection; The abnormal target stacks are transferred to the abnormal processing area, and there is no need to spend extra time separating the abnormal stacks from the normal ones.
  • this embodiment provides another stacking detection method, which, on the basis of the foregoing, focuses on describing a specific implementation of detecting the target stacking based on the initial number of items in the target stacking (that is, the number of items at the time of entry). Way.
  • the storage management method in this embodiment mainly includes the following steps S802 to S814:
  • Step S802 acquiring the initial quantity of the target stack to be put into storage.
  • the initial quantity of the target stack to be put into storage can be input by the warehouse staff from the terminal device, and then transmitted to the server by the terminal device, or the server can directly retrieve the target stack from the database storing the product information. information, and obtain the initial number of goods in the target stack from the goods information.
  • the warehouse staff wants to stack the detection target A, they can input the goods information of the target stack A from the terminal device (computer, mobile phone, tablet computer, etc.), or they can use other methods such as mechanical equipment, RFID ( Radio Frequency Identification, radio frequency identification), etc. to input the goods information, and the terminal device sends the goods information of the target stack A to the server.
  • the database in the server pre-stores the commodity information of the three target stacks B, C, and D. If you want to detect the target stack B, the server can directly retrieve the product information of the target stack B from the data.
  • the above-mentioned item information may include the item type and initial item quantity of the target stack.
  • Step S804 in the process of the target stack being conveyed on the conveying device, control multiple camera devices to collect multi-angle images of the target stack; Shoot the target stacks from different angles.
  • step S804 reference may be made to the related content of step S202, which will not be repeated here.
  • the multi-angle images of the target stack include side images of four sides and a top image of one top surface of the target stack.
  • step S806 the actual quantity of goods contained in the target stack is determined according to the multi-angle image of the target stack.
  • the server can calculate the actual number of items contained in the target stack. For example, the server can build a model of the target stack through the multi-angle image, and according to the established model, the actual number of items contained in the target stack can be determined; or, the server can obtain the area of each item in the multi-angle image and the entire target stack. The actual quantity of goods is calculated according to the above two areas.
  • the position of the goods in each multi-angle image (for example, the distance from the camera device) can also be analyzed, and the actual number of goods can be determined according to the position of the goods.
  • the multi-angle image captures the side images of 4 sides and the top image of 1 top surface of the target stack except the ground. It can be determined through steps A1-A2 that the target stack is in the target stack.
  • Step A1 take each side image and top image as the target image respectively, and mark in each target image the frame of goods contained on the stacking surface closest to the target camera device; wherein, the target camera device is a camera that collects the target image. equipment.
  • the target image in FIG. 9 is a side surface of the target stack, and the target stack is marked by the item frame in FIG. 9 , for example: the first area in FIG. 9 A total of 6 product boxes are marked, and a total of 9 product boxes are marked in the second area.
  • the item frame in the first area is not the item frame closest to the target imaging device, and the item frame in the second area is the item frame closest to the target imaging apparatus.
  • Step A2 Determine the actual number of goods contained in the target stack based on the goods frame contained in each marked target image.
  • the distribution of each layer of goods can be determined according to the number of layers of the goods, that is, the actual number of goods contained in the target stack can be calculated according to the marked goods frame.
  • the server can determine the distance of the goods in the goods box from the camera device according to the goods box, so as to determine the specific position of the goods in the goods box, and the server can determine the actual number of goods based on the positions of all goods in the target stack.
  • the distance between each item in the target stack and the camera device can be determined by marking all the multi-angle images of the target stack, and then the actual number of items contained in the target stack can be determined quickly and accurately. Calculate the actual number of items contained in the target stack.
  • Step S808 it is judged whether the actual quantity of goods is consistent with the initial quantity of goods. If they are consistent, go to step S812, if not, go to step S814.
  • the actual item quantity is consistent with the initial item quantity, it means that there is no problem with the quantity of items in the target stack. If other aspects do not need to be checked, it can be stored normally; if the actual item quantity is inconsistent with the initial item quantity, it means that the target stack If the number of goods in the stack is abnormal, subsequent inspections in other aspects may not be performed.
  • the stacking type or packaging of the target stack can also be detected, such as the detection items in the third embodiment, which will not be repeated here.
  • the above several stacking detection items can be used at the same time or not at the same time.
  • the stacking is normal stacking, and the target stacking is also abnormal stacking.
  • Step S812 control the conveying equipment to normally put the target stack into the warehouse.
  • Step S814 control the conveying device to convey the target stack to the exception handling area.
  • the target stack is an abnormal stack, it needs to be placed in a different area from the normal stack. Therefore, an area can be set as the exception handling area, and the exception handling area is responsible for storing the abnormal stack.
  • the abnormal stack can be conveyed by conveying equipment to the exception handling area. The abnormal stacking in this way will not be put into the warehouse, nor will it be placed together with the normal stacking, and the staff does not need to spend a lot of cost to separate the abnormal stacking from the normal stacking.
  • the server can send an early warning message to the warehouse staff, indicating that the target stacking is abnormal or the target stacking is stored in the abnormal processing area.
  • the early warning information carries the identification information of the target stack, and can also carry the image of the target stack (such as one or more of the above-mentioned multi-angle images). If the number of goods in the target stack is abnormal, it can also carry the target stack. the actual quantity of the item.
  • the above identification information may be a location identification, a serial number identification, or an abnormal cause identification, etc., and the abnormal target stack may be identified. After receiving the warning information, the staff can go to the exception handling area to view the target stack.
  • the target stacking can be manually or by other means (such as mechanical equipment, RFID, etc.) to perform the initial input of the quantity of goods, the target stacking can be transferred by the conveying device, and then passed through the warehouse gate or designated monitoring.
  • the camera device on the gantry scans the multi-angle images of the target stack, and counts the number of goods in the target stack to determine whether the target stack is abnormal.
  • This method can be quickly and accurately carried out by the server.
  • Stacking detection requires little manpower and time cost, high detection efficiency and detection success rate, and does not require high-cost high-density grating equipment. After detection, abnormal stacking and normal stacking are transferred to different areas. Workers do not need to spend a lot of money to separate abnormal stacks from normal stacks.
  • the product information x is manually input, and after the server receives the product information x, the Send the conveying instruction a to the conveying equipment, and the conveying equipment executes the conveying instruction a to convey the stack to the designated monitoring point. There are multiple camera devices at the designated monitoring point. After the conveying equipment transports the stack to the designated monitoring point, the server sends a photographing instruction b to the camera equipment. Angular image to server. The server receives multi-angle images and judges whether the stacking is abnormal according to the multi-angle images.
  • the server can archive the information of the stacking, send the conveying instruction c to the conveying device, and the conveying device executes the conveying instruction c to transport the stacking into the warehouse.
  • the server can mark and archive the abnormal area in the multi-angle image, send the conveying instruction d to the conveying equipment, the conveying equipment executes the conveying instruction d, and transports the stack to the abnormal processing area, and then the staff will manually deal with.
  • early warning information can also be sent to the terminal equipment of the staff, so that the staff can perform manual processing after receiving the early warning information.
  • the device is applied to a server, that is, the device can be installed on the server side, and the server is communicatively connected to a plurality of camera devices;
  • the device includes the following modules:
  • the first control module 1102 is configured to control a plurality of camera devices to collect multi-angle images of the target stack during the process of the target stack being conveyed on the conveying device; wherein, the plurality of camera devices are arranged on the conveying path of the target stack , shoot the target stacks from multiple different angles;
  • the second control module 1104 is configured to perform image analysis on the multi-angle images of the target stack, and determine whether the target stack is normally put into storage according to the analysis result.
  • the above-mentioned storage management device collects multi-angle images of the target stack by a plurality of camera devices during the conveying process of the target stack, performs image analysis on the multi-angle images of the target stack, and analyzes the multi-angle images of the target stack according to the analysis result. Determine whether the target stack is normally put into storage.
  • the server analyzes the multi-angle images of the target stack to determine whether it is normally put into storage, and then does not need to manually perform stacking inspection, which can save labor costs and time costs; image analysis and detection of multi-angle images can quickly , Accurately determine whether the target stack can be normally put into storage, thereby improving the detection efficiency and detection accuracy of stack detection.
  • the above-mentioned conveying device is provided with a photoelectric sensor, and the photoelectric sensor is configured to detect whether the target stack has reached the designated monitoring point; based on this, the above-mentioned first control module 1102 is also configured to detect that the target stack arrives at the designated monitoring point when the photoelectric sensor detects that the target stack reaches the designated monitoring point.
  • an image acquisition instruction is sent to multiple camera devices, so that the multiple camera devices collect multi-angle images of the target stack.
  • the above-mentioned conveying equipment includes a linear conveying line, and a plurality of camera devices are arranged on the gantry corresponding to the linear conveying line; the designated monitoring point is the monitoring point corresponding to the gantry.
  • the number of the above-mentioned multiple camera devices is five, wherein, four camera devices are respectively configured to photograph the four sides of the target stack located at the monitoring point, and one camera device is configured to photograph the top surface of the target stack. .
  • the above-mentioned conveying device includes a first conveying line and a second conveying line that constitute an L-shaped conveying line
  • the photoelectric sensor includes a first photoelectric sensor and a second photoelectric sensor
  • the first photoelectric sensor and the second photoelectric sensor are respectively arranged in the first and second photoelectric sensors.
  • the above-mentioned first control module 1102 is further configured to: when the first photoelectric sensor detects that the target stack arrives at the designated monitoring point on the first conveying line, send a message to the first group of camera equipment. Image acquisition instruction; when the second photoelectric sensor detects that the target stack reaches the designated monitoring point on the second conveying line, an image acquisition instruction is sent to the second group of camera devices.
  • the number of the above-mentioned multiple camera devices is 5, wherein, the first group of camera devices includes 3 camera devices, which are respectively configured to photograph the two sides of the target stack located at the designated monitoring point on the first conveying line. and the top surface of the target stack; the second group of camera devices includes two camera devices, which are respectively configured to photograph the other two sides of the target stack located at the designated monitoring point on the second conveying line.
  • the above-mentioned second control module 1104 is further configured to: perform at least one of the following detections on the multi-angle images of the target stack through a pre-trained image analysis model: detection of the quantity of goods, detection of the shape of the stack, and detection of the deformation of the packaging of goods. , Product packaging damage detection and product packaging stain detection; determine whether the target stacking is abnormal according to the detection results.
  • the above-mentioned stack type detection includes at least one of the following: detecting the inclination of the target stack; detecting the placement relationship between the goods in the target stack and the tray; detecting the up and down of the goods in the target stack Placement relationship; detect the distance between items in the target stack.
  • the above-mentioned apparatus further includes: an information acquisition module, which is connected to the first control module 1102 or the second control module 1104, and the information acquisition module is configured to: acquire the target heap to be stored in the library
  • the second control module 1104 is further configured to: determine the actual number of goods contained in the target stack according to the multi-angle image of the target stack; determine whether the actual number of goods is consistent with the initial number of goods; , to determine the abnormality of the target stacking.
  • the above-mentioned multi-angle images include side images of 4 side surfaces and a top surface image of one top surface of the target stack; correspondingly, the second control module 1104 is further configured to: separate each side image and the top surface of the stack.
  • the image is used as a target image, and each target image is marked with the goods frame contained on the stacking surface closest to the target camera device; wherein, the target camera device is the camera device that collects the target image;
  • the item box determines the actual number of items contained in the target stack.
  • the above-mentioned second control module 1104 is further configured to: if the analysis result is that the target stack is abnormal, control the transfer device to transfer the target stack to the abnormal processing area; if the analysis result is that the target stack is normal, control the transfer device to transfer the target stack.
  • the stacking is normally put into storage.
  • An embodiment of the present disclosure provides a warehouse management system, including a server, a conveying device communicatively connected to the server, and a plurality of camera devices; the conveying device is configured to convey target stacks under the control of the server; the server is configured to: In the process of stacking on the conveying equipment, control multiple camera devices to collect multi-angle images of the target stack; and perform image analysis on the multi-angle images of the target stack, and determine whether the target stack is normally loaded according to the analysis results. Library; multiple camera devices are set on the conveying path of the target stack, and the target stack is photographed from multiple different angles.
  • the above-mentioned conveying equipment includes a linear conveying line, and a plurality of imaging devices are arranged on the gantry corresponding to the linear conveying line;
  • the equipment is arranged on the first gantry corresponding to the L-shaped conveying line, and the second group of imaging devices among the plurality of imaging devices is arranged on the second gantry corresponding to the L-shaped conveying line.
  • An embodiment of the present disclosure provides an electronic system, the electronic system includes: a processing device and a storage device; a computer program is stored on the storage device, and the computer program executes the above-mentioned storage management method when the processed device runs.
  • Embodiments of the present disclosure further provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is run by a processing device, the steps of the above-mentioned storage management method are executed.
  • the storage management method, device, warehouse management system, and computer program product of an electronic system provided by the embodiments of the present disclosure include a computer-readable storage medium storing program codes, and the instructions included in the program codes can be used to execute the foregoing method embodiments.
  • the specific implementation can refer to the method embodiment, which is not repeated here.
  • the terms “installed”, “connected” and “connected” should be understood in a broad sense, for example, it may be a fixed connection or a detachable connection , or integrally connected; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or the internal communication between the two components.
  • installed should be understood in a broad sense, for example, it may be a fixed connection or a detachable connection , or integrally connected; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or the internal communication between the two components.
  • the functions, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer-readable storage medium.
  • the computer software products are stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of various embodiments of the present disclosure.
  • the aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .
  • the server determines whether the target stacking is normally put into storage by performing image analysis on the multi-angle images of the target stack, thereby eliminating the need for manual stacking detection, which can save labor costs and time costs;
  • the angle image can quickly and accurately determine whether the target stack can be stored normally, thereby improving the detection efficiency and detection accuracy of stack detection, and ensuring the quality of the stack in storage.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Human Resources & Organizations (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Finance (AREA)
  • Accounting & Taxation (AREA)
  • Multimedia (AREA)
  • Warehouses Or Storage Devices (AREA)

Abstract

一种入库管理方法、装置、仓库管理***和电子***，其中，该方法包括：该方法应用于服务器，该服务器与多个摄像设备通信连接；该方法包括：在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；其中，多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛；对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库。该方式无需人工进行堆垛检测，可以节约人力成本和时间成本；通过图像分析检测多角度图像可以快速、准确地确定目标堆垛是否正常入库，保证了入库堆垛的质量。

Description

入库管理方法、装置、仓库管理***和电子***
相关申请的交叉引用
本公开要求于2020年11月23日提交中国专利局的申请号为202011325601.1、名称为“入库管理方法、装置、仓库管理***和电子***”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。
技术领域
本公开涉及堆垛入库技术领域，尤其是涉及一种入库管理方法、装置、仓库管理***和电子***。
背景技术
传统的堆垛入库主要依靠对射光栅对堆垛的垛型进行检测，货物的数量盘点复核和堆垛外形检测主要依靠人工观察。其中，一般方法是通过由对射光栅组成的安全光幕发射红外线形成保护光幕，对货物进行二维成像，对成像结果进行分析，最终确定堆垛的垛型是否合格。但是这种方法需要耗费较高的人力和时间成本，而且检测效率及检测准确率均较低。
公开内容
有鉴于此，本公开的目的在于提供一种入库管理方法、装置、仓库管理***和电子***，以节约人力和时间资源，提高检测效率和检测准确率。
第一方面，本公开实施例提供了一种入库管理方法，该方法应用于服务器，该服务器与多个摄像设备通信连接；该方法包括：在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；其中，多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛；对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库。
可选的，上述传送设备上设置有光电传感器，该光电传感器配置成检测目标堆垛是否到达指定监测点；上述控制多个摄像设备采集目标堆垛的多角度图像，包括：当光电传感器监测到目标堆垛到达指定监测点时，向多个摄像设备发送图像采集指令，以使多个摄像设备采集目标堆垛的多角度图像。
可选的，上述传送设备包括直线型输送线，多个摄像设备设置在直线型输送线对应的门架上；上述指定监测点为门架对应的监测点。
可选的，多个摄像设备的数量为5个，其中，4个摄像设备分别配置成拍摄位于监测点的目标堆垛的四个侧面，1个摄像设备配置成拍摄目标堆垛的顶面。
可选的，上述传送设备包括构成L型输送线的第一输送线和第二输送线，上述光电传感器包括第一光电传感器和第二光电传感器，第一光电传感器和第二光电传感器分别设置在第一输送线和第二输送线上，多个摄像设备中的第一组摄像设备设置在第一输送线的第一门架上，多个摄像设备中的第二组摄像设备设置在第二输送线的第二门架上；上述当光电传感器监测到目标堆垛到达指定监测点时，向多个摄像设备发送图像采集指令，包括：当第一光电传感器监测到目标堆垛到达第一输送线上的指定监测点时，向第一组摄像设备发送图像采集指令；当第二光电传感器监测到目标堆垛 到达第二输送线上的指定监测点时，向第二组摄像设备发送图像采集指令。
可选的，多个摄像设备的数量为5个，其中，第一组摄像设备包括3个摄像设备，分别配置成拍摄位于第一输送线上的指定监测点的目标堆垛的两个侧面和目标堆垛的顶面；第二组摄像设备包括2个摄像设备，分别配置成拍摄位于第二输送线上的指定监测点的目标堆垛的另两个侧面。
可选的，上述对目标堆垛的多角度图像进行图像分析，包括：通过预先训练好的图像分析模型，对目标堆垛的多角度图像进行以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测；根据检测结果确定目标堆垛是否异常。
可选的，上述图像分析包括垛型检测，上述垛型检测包括以下至少之一：检测目标堆垛的倾斜程度；检测目标堆垛中的货品与托盘的摆放关系；检测目标堆垛中的货品上下摆放关系；检测目标堆垛中的货品间距。
可选的，上述对目标堆垛的多角度图像进行图像分析之前，方法还包括：获取待入库的目标堆垛的初始货品数量；
上述对目标堆垛的多角度图像进行图像分析，包括：根据目标堆垛的多角度图像确定目标堆垛中包含的实际货品数量；判断实际货品数量与初始货品数量是否一致；如果不一致，确定目标堆垛异常。
可选的，多角度图像包括目标堆垛的4个侧面的侧面图像和1个顶面的顶面图像；上述根据目标堆垛的多角度图像确定目标堆垛中包含的实际货品数量的步骤，包括：分别将每个侧面图像和顶面图像作为目标图像，在每个目标图像中标注出距离目标摄像设备最近的堆垛表面包含的货品框；其中，目标摄像设备为采集目标图像的摄像设备；基于标注后的每个目标图像包含的货品框确定目标堆垛中包含的实际货品数量。
可选的，上述根据分析结果确定是否将目标堆垛正常入库，包括：如果分析结果为目标堆垛异常，控制传送设备将目标堆垛传送至异常处理区；如果分析结果为目标堆垛正常，控制传送设备将目标堆垛正常入库。
第二方面，本公开实施例还提供一种入库管理装置，该装置应用于服务器，该服务器与多个摄像设备通信连接；该装置包括：第一控制模块，配置成在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛；第二控制模块，配置成对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库。
可选的，所述第二控制模块还配置成：通过预先训练好的图像分析模型，对所述目标堆垛的多角度图像进行以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测；根据检测结果确定所述目标堆垛是否异常。
可选的，所述装置还包括：信息获取模块，配置成：获取待入库的目标堆垛的初始货品数量；所述第二控制模块还配置成：根据所述目标堆垛的多角度图像确定所述目标堆垛中包含的实际货品数量；判断所述实际货品数量与所述初始货品数量是否一致；如果不一致，确定所述目标堆垛异常。
可选的，所述第二控制模块还配置成：如果分析结果为所述目标堆垛异常，控制所述传送设备将所述目标堆垛传送至异常处理区；如果分析结果为所述目标堆垛正常，控制所述传送设备将所述目标堆垛正常入库。
第三方面，本公开实施例还提供一种仓库管理***，包括服务器、与服务器通信连接的传送设备和多个摄像设备；传送设备配置成在服务器的控制下传送目标堆垛；服务器配置成：在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；以及对目标堆垛 的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库；多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛。
在本公开另一实施例中，上述传送设备包括直线型输送线，多个摄像设备设置在直线型输送线对应的门架上；或者，传送设备包括L型输送线，多个摄像设备中的第一组摄像设备设置在L型输送线对应的第一门架上，多个摄像设备中的第二组摄像设备设置在L型输送线对应的第二门架上。
第四方面，本公开实施例还提供一种电子***，电子***包括：处理设备和存储装置；存储装置上存储有计算机程序，计算机程序在被处理设备运行时执行上述的入库管理方法。
第五方面，本公开实施例还提供一种计算机可读存储介质，计算机可读存储介质上存储有计算机程序，计算机程序被处理设备运行时执行如上述入库管理方法的步骤。
本公开实施例带来了以下有益效果：
本公开实施例提供的一种入库管理方法、装置、仓库管理***和电子***，通过在目标堆垛传送过程中，由多个摄像设备采集目标堆垛的多角度图像，对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将该目标堆垛正常入库。该方式中，服务器通过对目标堆垛的多角度图像进行图像分析，确定其是否正常入库，进而无需人工进行堆垛检测，可以节约人力成本和时间成本；通过图像分析检测多角度图像可以快速、准确地确定目标堆垛是否可以正常入库，从而提升堆垛检测的检测效率和检测准确率。
本公开的其他特征和优点将在随后的说明书中阐述，或者，部分特征和优点可以从说明书推知或毫无疑义地确定，或者通过实施本公开的上述技术即可得知。
为使本公开的上述目的、特征和优点能更明显易懂，下文特举较佳实施例，并配合所附附图，作详细说明如下。
附图说明
为了更清楚地说明本公开的技术方案，下面将对其中所需要使用的附图作简单地介绍，应当理解，以下附图仅示出了本公开的某些实现方式，因此不应被看作是对范围的限定，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它相关的附图。
图1为本公开实施例提供的一种电子***的结构示意图；
图2为本公开实施例提供的一种入库管理方法的流程图；
图3为本公开实施例提供的一种摄像设备设置方式示意图；
图4为本公开实施例提供的另一种摄像设备设置方式示意图；
图5为本公开实施例提供的另一种入库管理方法的流程图；
图6为本公开实施例提供的一种存在倾斜散落风险的目标堆垛的示意图；
图7为本公开实施例提供的一种目标堆垛的垛型异常的示意图；
图8为本公开实施例提供的另一种入库管理方法的流程图；
图9为本公开实施例提供的一种目标图像的示意图；
图10为本公开实施例提供的一种入库管理方法的整体流程示意图；
图11为本公开实施例提供的一种入库管理装置的结构示意图。
具体实施方式
为使本公开实施例的目的、技术方案和优点更加清楚，下面将结合附图对本公开的技术方案进行清楚、完整地描述，显然，所描述的实施例是本公开一部分实施例，而不是全部的实施例。基于 本公开中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本公开保护的范围。
目前，堆垛入库时主要由人工进行货物的数量盘点和外形检测，人工盘点数量和检测外形的效率和准确率均不高。具体而言，通过人工盘点数量需要较高的人力和时间成本，效率较低，连续性也较差，且人工盘点数量无法胜任货物过多情况下的大仓库检测；人工检测堆垛外形的准确率较低，且难以发现堆垛表面的全部污迹，也无法保证入库堆垛的质量全部符合要求；此外，通过光幕对货物进行二维成像的局限性较大，需要高成本的高密度光栅设备才可以获得较高的检测精确度；不合格的堆垛入库后再取出需要花费大量成本，处理不及时可能会造成额外损失，处理不合格的堆垛的时间成本和人力成本也很高。
为改善上述问题至少之一，本公开实施例提供了一种入库管理方法、装置、仓库管理***和电子***，该技术可以应用于手机、电脑、服务器等需要进行堆垛检测的电子***中。
为便于对本实施例进行理解，下面对本公开实施例所公开的一种入库管理方法、装置、仓库管理***和电子***进行详细介绍。
首先，参照图1所示的电子***100的结构示意图。该电子***可以用于实现本公开实施例的入库管理方法和装置，以及仓库管理***。
如图1所示的一种电子***的结构示意图，电子***100包括一个或多个处理设备102、一个或多个存储装置104、输入装置106、输出装置108以及一个或多个图像采集设备110，这些组件通过总线***112和/或其它形式的连接机构(未示出)互连。应当注意，图1所示的电子***100的组件和结构只是示例性的，而非限制性的，根据需要，电子***也可以具有其他组件和结构。
处理设备102可以为服务器、智能终端，或者是包含中央处理单元(CPU)或者具有数据处理能力和/或指令执行能力的其它形式的处理单元的设备，可以对电子***100中的其它组件的数据进行处理，还可以控制电子***100中的其它组件以执行入库管理的功能。
存储装置104可以包括一个或多个计算机程序产品，计算机程序产品可以包括各种形式的计算机可读存储介质，例如易失性存储器和/或非易失性存储器。易失性存储器例如可以包括随机存取存储器(RAM)和/或高速缓冲存储器(cache)等。非易失性存储器例如可以包括只读存储器(ROM)、硬盘、闪存等。在计算机可读存储介质上可以存储一个或多个计算机程序指令，处理设备102可以运行程序指令，以实现下文的本公开实施例中(由处理设备实现)的客户端功能以及/或者其它期望的功能。在计算机可读存储介质中还可以存储各种应用程序和各种数据，例如应用程序使用和/或产生的各种数据等。
输入装置106可以是用户用来输入指令的装置，并且可以包括键盘、鼠标、麦克风和触摸屏等中的一个或多个。
输出装置108可以向外部(例如，用户)输出各种信息(例如，图像或声音)，并且可以包括显示器、扬声器等中的一个或多个。
图像采集设备110(也可以称为摄像设备，例如摄像头等装置)可以获取目标堆垛的多角度图像，并且将采集到的图像存储在存储装置104中以供其它组件使用。
示例性地，用于实现根据本公开实施例的入库管理方法、装置、仓库管理***和电子***中的各器件可以集成设置，也可以分散设置，诸如将处理设备102、存储装置104、输入装置106和输出装置108集成设置于一体，而将图像采集设备110设置于可以采集到图像的指定位置。当上述电子***中的各器件集成设置时，该电子***可以被实现为诸如相机、智能手机、平板电脑、计算机、 车载终端等智能终端。
本实施例提供了一种入库管理方法，该方法应用于服务器，也即，该方法可由服务器执行，该服务器可以是仓库管理***中的一个控制设备，与仓库管理***中的传送设备、摄像设备等通信连接，本公开实施例中的摄像设备可以为摄像头、监控器等可以进行图像采集的设备，该摄像设备通常设置在堆垛传送路径的指定监测点。本实施例的摄像设备可以有多个，为了能够清晰地采集堆垛的多角度图像，多个摄像设备可以在指定监测点的多个不同方位上。这里需要说明的是，摄像设备的数量和安放位置根据图像采集的需要确定，可以保证摄像设备尽可能多地采集堆垛不同面的图像即可。
基于上述描述，参见图2所示的一种入库管理方法的流程图，该入库管理方法主要包括如下步骤S202～步骤S204：
步骤S202，在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集该目标堆垛的多角度图像。其中，多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄该目标堆垛。
有需要入库的货物时，通常将该货物放在托盘上码成一个堆垛，再将该堆垛放在传送设备上传送入库。为了便于描述，将放在传送设备上需要入库的堆垛称为目标堆垛。
传送设备可以为带有传送带的设备或者传送小车等。目标堆垛在传送过程中，目标堆垛会沿着堆垛传送路径进行传输，由于堆垛传送路径的指定监测点处设置有多个摄像设备，在实际应用中，可以在一个指定监测点处设置多个朝向不同的摄像设备，也可以在多个指定监测点处分别设置至少一个摄像设备，在此不进行限制。通过设置多个摄像设备，可以在目标堆垛进行传送时采集目标堆垛的多角度图像。
多角度图像包括从多个不同角度拍摄目标堆垛得到的多个图像；通过不同角度可以拍摄目标堆垛的多个可视面对应的图像，可视面即在传送过程中人眼可视的表面。例如：假设目标堆垛为立方体，共有6个表面，除了下表面在传送过程中人眼不可视，其余5个表面(上表面、前表面、后表面、左表面和右表面)均为可视面。
这里需要说明的是，多个摄像设备可以在不同时间采集目标堆垛处于不同位置的多角度图像，每个摄像设备采集的图像对应的目标堆垛表面可以相同，也可以不同。为了保证后续检测的准确率，可以采集目标堆垛的全部可视面对应的图像。
例如，目标堆垛在第3秒处于堆垛传送路径的2米处，摄像设备1可以采集目标堆垛上表面的图像；目标堆垛在第10秒处于堆垛传送路径的15米处，摄像设备2可以采集目标堆垛右表面的图像；目标堆垛在第20秒处于堆垛传送路径的30米处，摄像设备3也可以采集目标堆垛上表面的图像。
在目标堆垛在传送设备上进行传送的过程中，服务器可以在目标堆垛当前所在位置与摄像设备所在位置比较接近时，控制摄像设备开启采集功能，进而采集到该目标堆垛的多个不同角度的图像。
目标堆垛当前所在位置可以根据传送设备的传送速度确定，也可以根据传送设备上安装的检测传感器反馈的检测信息等确定，对此本公开实施例不进行限定。
步骤S204，对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库。
服务器可以对多角度图像进行图像分析，该图像分析可以包括目标堆垛包含的货品是否摆放正常，货品数量是否正确，以及货品的外包装是否有破损或者有污渍等分析项目中的一个或多个。即 图像分析包括以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测。图像分析的手段可以包括图像检测、图像识别、图像分割等，通过对多角度图像进行分析，可以确定多角度图像中包含的货品的轮廓，以及货品数量等。然后，再根据分析结果确定是否将目标堆垛正常入库，通常，需要分析的项目都正常或正确，则将目标堆垛正常入库，否则进行异常处理，如将目标堆垛传送到指定的异常处理区域，以供后续处理。
例如，在对目标堆垛的多角度图像进行货品数量检测时，可以根据托盘的尺寸计算堆垛的长宽高，并计算堆垛的体积，然后根据堆垛中码放的单个货品的体积，进一步计算对堆垛中码放的货品数量。还可以将多角度图像输入到预先训练好的数量检测模型中，由数量检测模型输出该目标堆垛的数量。
上述方法通过在目标堆垛传送过程中，由多个摄像设备采集目标堆垛的多角度图像，对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将该目标堆垛正常入库。该方式中，服务器通过对目标堆垛的多角度图像进行图像分析，确定其是否正常入库，进而无需人工进行堆垛检测，可以节约人力成本和时间成本；通过图像分析检测多角度图像可以快速、准确地确定目标堆垛是否可以正常入库，从而提升堆垛检测的检测效率和检测准确率。
为了能够比较精准的控制摄像设备进行图像采集，本实施例的传送设备上设置有光电传感器，该光电传感器配置成检测目标堆垛是否到达指定监测点；相应地，上述控制多个摄像设备采集目标堆垛的多角度图像，包括：当上述光电传感器监测到目标堆垛到达指定监测点时，向多个摄像设备发送图像采集指令，以使多个摄像设备采集目标堆垛的多角度图像。具体地，本实施例的服务器可以接收光电传感器在监测到目标堆垛到达指定监测点时发送的提示信息，并根据该提示信息向多个摄像设备发送图像采集指令。这种通过传送设备上的光电传感器检测堆垛是否到达指定监测点，进而触发摄像设备进行图像采集的方式，可以比较准确地把控触发摄像设备的时机，进而保证采集的图像质量。
对于直线型传送线的传送设备，上述多个摄像设备可以设置在直线型输送线对应的门架上；上述指定监测点为该门架对应的监测点。通常，为了简化摄像设备的安装和布设成本，可以在直线型输送线对应的一个门架上安装多个摄像设备，例如：多个摄像设备的数量为5个，其中，4个摄像设备分别配置成拍摄位于该监测点的目标堆垛的四个侧面，1个摄像设备用于拍摄该目标堆垛的顶面。作为一种可能的实施方式，4个摄像设备中的每个摄像设备均与所拍摄侧面的距离在第一预设距离区间(例如：0.9米至1.2米，可选的，距离设置为1米)内，与目标堆垛的顶面的距离在第二预设距离区间(例如：0.1米至0.3米，可选的，距离设置为0.2米)内，与水平面对应的拍摄角度在第一预设角度区间(例如：40度至50度，可选的，角度设置为45度)内。而另一个拍摄目标堆垛顶面的摄像设备，与目标堆垛的顶面的距离在第二预设距离区间内，其拍摄角度垂直指向该目标堆垛。
对于包含L型输送线的传送设备，该传送设备包括构成L型输送线的第一输送线和第二输送线，上述光电传感器包括第一光电传感器和第二光电传感器，其中，第一光电传感器和第二光电传感器分别设置在第一输送线和第二输送线上，多个摄像设备中的第一组摄像设备设置在第一输送线的第一门架上，多个摄像设备中的第二组摄像设备设置在第二输送线的第二门架上；相应地，上述当光电传感器监测到目标堆垛到达指定监测点时，向多个摄像设备发送图像采集指令，包括：当第一光电传感器监测到目标堆垛到达第一输送线上的指定监测点时，向第一组摄像设备发送图像采集指令；当第二光电传感器监测到目标堆垛到达第二输送线上的指定监测点时，向第二组摄像设备发送图像采集指令。
具体实现时，上述第一光电传感器在监测到目标堆垛时，向服务器发送提示信息，以提示目标堆垛已运送至第一门架对应的第一监测点；上述第二光电传感器在监测到目标堆垛时，向服务器发送提示信息，以提示目标堆垛已运送至第二门架对应的第二监测点。为了区分两个提示信息，提示信息可以携带光电传感器的标识信息，或者携带预先规定的信息标识符，以起到提示目标堆垛当前所在位置的目的。当服务器接收到的提示信息为目标堆垛已运送至第一门架对应的第一监测点时，向第一组摄像设备分别发送图像采集指令；当接收到提示信息为目标堆垛已运送至第二门架对应的第二监测点时，向第二组摄像设备分别发送图像采集指令。
对于包含L型输送线的传送设备，上述多个摄像设备的数量也可以为5个，其中，第一组摄像设备包括3个摄像设备，分别用于拍摄位于第一输送线上的指定监测点的目标堆垛的两个侧面和目标堆垛的顶面；第二组摄像设备包括2个摄像设备，分别用于拍摄位于第二输送线上的指定监测点的目标堆垛的另两个侧面。
其中，第一组摄像设备中的拍摄两个侧面的摄像设备：与所拍摄侧面的距离在第三预设距离区间内，与目标堆垛的顶面的距离在第四预设距离区间内，与水平面对应的拍摄角度在第二预设角度区间内；第一组摄像设备中拍摄目标堆垛顶面的摄像设备与该顶面的距离在第四预设距离区间内，拍摄角度垂直指向目标堆垛；第二组摄像设备中的摄像设备：与所拍摄侧面的距离在第三预设距离区间内，与目标堆垛的顶面的距离在第四预设距离区间内，与水平面对应的拍摄角度在第二预设角度区间内。上述距离区间和角度区间可以参考直线型输送线中的参数区间，这里不再赘述。上述第一预设距离区间、第二预设距离区间、第三预设距离区间和第四预设距离区间彼此之间可相同也可不同，第一预设角度区间和第二预设角度区间也可相同也可不同，具体可根据实际情况灵活设置，在此不进行限制。
参见图3所示的一种摄像设备设置方式示意图，摄像设备1至摄像设备5可以为摄像头，在实际应用中，可以在堆垛传送路上设置有仓库的指定门架，摄像设备1至摄像设备5设置在该指定门架上，一种具体示例中，摄像设备1至摄像设备4分别距离托盘边缘1米，高度可以比目标堆垛的最高通行高度高0.2米，角度可以和水平面呈45度夹角向下。摄像设备5可以设置在托盘和堆垛的中央正上方，高度比目标堆垛的最高通行高度高1米，垂直向下拍摄。
参见图4所示的另一种摄像设备设置方式示意图，摄像设备1、摄像设备2、摄像设备5设置在第一门架上，分别采集目标堆垛的2个侧面和顶面的图像，一种具体示例中，摄像设备1、2距离托盘和堆垛1米，高度为目标堆垛的最高通行高度的一半，角度保持水平，摄像设备5可以设置在托盘和堆垛的中央正上方，高度比目标堆垛的最高通行高度高1米，角度垂直向下。
托盘通过第一门架后随“L型”传送带继续运行，堆垛相对于地面的绝对方向不变，传送带运行方向顺时针旋转90度，摄像设备3、摄像设备4设置在第二门架上，分别采集目标堆垛的另外2个侧面的多角度图像，摄像设备3、摄像设备4的位置与摄像设备1、摄像设备2类似。此外，摄像设备1也可以设置在第二门架上，这里不做限定。上述两种设置方式，可以通过设置较少的门架采集目标堆垛的多角度图像，除此以外，还可以将摄像设备设置在更多数量的门架上。
在前述基础上，本实施例提供了另一种入库管理方法，该方法应用于上述服务器，重点描述基于图像分析模型进行图像分析的方式，参见图5所示的另一种入库管理方法的流程图，该方法主要包括以下步骤S502～步骤S510：
步骤S502，在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛；步 骤S502可以参照步骤S202的相关内容，这里不再赘述。
步骤S504，通过预先训练好的图像分析模型，对目标堆垛的多角度图像进行以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测。
一般来说，堆垛为立方体结构，共有6个表面，目标堆垛在传送设备上通过托盘进行运输，由于目标堆垛的下表面紧贴托盘，无法对下表面采集图像。因此，为了全方位检测目标堆垛是否异常，可以通过摄像设备采集5个表面对应的图像，具体为目标堆垛的4个侧面的侧面图像和1个顶面的顶面图像。通过预先训练好的图像分析模型对这五个图像进行图像分析。该图像分析模型可以基于各个堆垛对应的图像样本进行训练，具体训练过程可以参考相关模型训练的方式，这里不再详述。
通常，对于单一货品的仓库，每个堆垛的货品数量可以是预先规定的数量，该数量通常固定不变，因此上述货品数量检测可以基于目标堆垛的多角度图像统计出该目标堆垛的实际货品数量，如果该实际货品数量与规定的数量一致，则该目标堆垛的货品数量是正常的。
本公开实施例中的垛型即堆垛的形状，服务器可以根据目标堆垛的多角度图像确定目标堆垛的垛型。通过垛型可以看出目标堆垛是否存在摆放松散、错位摆放等问题，从而确定目标堆垛是否出现倾斜散落风险。参见图6所示的一种存在倾斜散落风险的目标堆垛的示意图，可以看出，图6中上方的目标堆垛摆放整齐，不存在倾斜散落风险；而下方的目标堆垛摆放散乱，存在倾斜散落风险。基于此，为了能够较为准确的判别出垛型是否异常，上述垛型检测可以包括以下至少之一：
(1)检测目标堆垛的倾斜程度；
目标堆垛的倾斜程度可以由堆垛的侧面图像确定，如果侧面图像中堆垛的侧棱不再是一条直线，而是多个小线段组成，则说明堆垛倾斜，相邻小线段之间的距离越大，说明倾斜程度越大，此时堆垛容易倾斜散落的风险越大，视为堆垛异常。参见图7所示的目标堆垛的垛型异常的示意图，其中，图7中的3)所示的垛型的倾斜程度较大。
(2)检测目标堆垛中的货品与托盘的摆放关系；例如，货品是否超出托盘等。如图7中1)所示的商品超出托盘的一种情况。
(3)检测目标堆垛中的货品上下摆放关系，例如是否上宽下窄；参见图7中的4)所示的垛型上宽下窄的一种情况。
(4)检测目标堆垛中的货品间距。图7中6)所示的堆垛悬空，示意出了货品间距较大的一种情况。
除了上述不规则之外，图7所示的不符合垛型要求的垛型还包括2)不规则码放的示意图和5)上重下轻的示意图，其中，对于上重下轻的垛型而言，除了基于堆垛下部箱体的变形情况推测之外，还可以通过检测堆垛上部分的货品类型和下部分的货品类型推测。不同的货品类型，其单件箱体的重量通常不同，如果较重的货品位于堆垛上部分，容易造成下部分货品的积压变形和包装损坏，也容易在传送过程中发生掉落。
如果目标堆垛不符合至少一项预设的垛型要求，则说明该目标堆垛存在码放问题，从而可以认定为异常堆垛。本公开实施例提供的上述方法，从目标堆垛的多角度图像中确定目标堆垛的垛型、数量以及包装是否有损等，可以检测出异常堆垛。
此外需要说明的是，上述几种堆垛检测方法可以同时使用，也可以不同时使用，一般来说，只要有一种检测方法将目标堆垛认定为异常堆垛，即使其他的检测方法认定该目标堆垛为正常堆垛，该目标堆垛也为异常堆垛。
步骤S506，根据检测结果确定目标堆垛是否异常。如果正常，执行步骤S508；如果异常，执行步骤S510。
步骤S508，控制传送设备将目标堆垛正常入库。
步骤S510，控制传送设备将目标堆垛传送至异常处理区。
如果目标堆垛异常，则不能将该目标堆垛传送至仓库内，需要通过传送设备将该目标堆垛传送至异常处理区。异常处理区为异常的目标堆垛的放置区域，目标堆垛传送至异常处理区后，工作人员可以在异常处理区对该目标堆垛进行检查。
在确定目标堆垛异常之后，服务器可以向仓库工作人员发送预警信息，提示目标堆垛异常或者目标堆垛存入异常处理区，例如：如果目标堆垛异常，向指定的终端设备发送预警信息；其中，预警信息携带有目标堆垛的标识信息，也可以携带目标堆垛的图像(如上述多角度图像中的一个或多个)，如果目标堆垛的货品数量异常，还可以携带目标堆垛的实际货品数量。
其中，上述标识信息可以为位置标识、序号标识或者异常原因标识等，可以对异常的目标堆垛进行标识即可。工作人员在接收到预警信息之后，可以前往异常处理区对目标堆垛进行查看。
而对于正常的目标堆垛，则可以正常通过传送设备入库，从而实现正常的目标堆垛和异常的目标堆垛分隔放置，无需工作人员手动将异常的目标堆垛和正常的目标堆垛分离。
上述方法，由多个摄像设备采集目标堆垛的多角度图像，将目标堆垛的多角度图像输入图像分析模型进行货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测或货品包装污迹检测，如果目标堆垛异常，服务器控制传送设备将目标堆垛传送至异常处理区。该方式无需人工进行堆垛检测，可以节约人力成本和时间成本；通过图像分析检测多角度图像可以快速、准确地确定目标堆垛是否异常，从而提升堆垛检测的检测效率和检测准确率；将异常的目标堆垛传送至异常处理区，无需花费额外的时间将异常的堆垛和正常的堆垛分离。
为了便于对入库的货品进行管理，堆垛入库时，通常会录入需要入库的堆垛上的货品的相关信息，例如：货品类型和货品数量等。基于此，本实施例提供了另一种堆垛检测方法，该方法在前述基础上，重点描述基于目标堆垛的初始货品数量(即录入时的货品数量)检测目标堆垛的一种具体实施方式。如图8所示的另一种入库管理方法的流程图，本实施例中的入库管理方法主要包括如下步骤S802～步骤S814：
步骤S802，获取待入库的目标堆垛的初始货品数量。
待入库的目标堆垛的初始货品数量可以由仓库的工作人员从终端设备输入，然后由终端设备传输至服务器中，或者服务器直接从存储有货品信息的数据库中调取该目标堆垛的货品信息，从货品信息中得到该目标堆垛的初始货品数量。
例如：仓库的工作人员想要将检测目标堆垛A，可以从终端设备(电脑、手机、平板电脑等设备)上输入目标堆垛A的货品信息，也可以通过其他方式如机械设备、RFID(Radio Frequency Identification，射频识别)等进行货品信息输入，终端设备将目标堆垛A的货品信息发送至服务器中。或者，服务器中的数据库预先存储有三个目标堆垛B、C、D的货品信息。如果想要对目标堆垛B进行检测，服务器可以直接从数据中调取目标堆垛B的货品信息。上述货品信息可以包括目标堆垛的货品类型和初始货品数量。
步骤S804，在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；其中，多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛。步骤S804可参考步骤S202的相关内容，在此不再赘述。
本实施例中，目标堆垛的多角度图像包括目标堆垛的4个侧面的侧面图像和1个顶面的顶面图像。
步骤S806，根据目标堆垛的多角度图像确定目标堆垛中包含的实际货品数量。
通过目标堆垛的多角度图像，服务器可以计算出目标堆垛中包含的实际货品数量。例如：服务器可以通过多角度图像建立目标堆垛的模型，根据建立的模型可以确定目标堆垛中包含的实际货品数量；或者，服务器可以获取多角度图像中每个货品的面积以及整个目标堆垛的面积，根据上述两个面积计算实际货品数量。
除了上述两种方式之外，还可以分析每一个多角度图像中的货品的位置(例如距离摄像设备的距离)，根据货品的位置确定实际货品数量。假设目标堆垛为立方体结构，多角度图像采集了除地面之外的目标堆垛的4个侧面的侧面图像和1个顶面的顶面图像，可以通过步骤A1-步骤A2确定目标堆垛中包含的实际货品数量：
步骤A1，分别将每个侧面图像和顶面图像作为目标图像，在每个目标图像中标注出距离目标摄像设备最近的堆垛表面包含的货品框；其中，目标摄像设备为采集目标图像的摄像设备。
分别将目标堆垛的每个侧面图像和顶面图像作为目标图像，对目标图像进行图像分析。参见图9所示的一种目标图像的示意图，图9中的目标图像为目标堆垛的一个侧面，在图9中通过货品框对目标堆垛进行标注，例如：图9中的第一区域共标注了6个货品框，第二区域共标注了9个货品框。其中，第一区域的货品框并不是距离目标摄像设备最近的货品框，而第二区域的货品框是距离目标摄像设备最近的货品框。
步骤A2，基于标注后的每个目标图像包含的货品框确定目标堆垛中包含的实际货品数量。
在完成对目标堆垛的全部目标图像的标注之后，可以根据货品的层数确定每一层货品的分布，即根据标注的货品框计算目标堆垛包含的实际货品数量。例如：服务器可以根据货品框确定货品框中的货品距离摄像设备的距离，从而确定货品框中的货品的具***置，服务器可以基于目标堆垛中所有货品的位置确定实际货品数量。
该方式中，可以通过对目标堆垛的全部多角度图像进行标注的形式，确定目标堆垛中每一个货品距离摄像设备的距离，然后确定出目标堆垛包含的实际货品数量，可以快速、准确地计算目标堆垛包含的实际货品数量。
步骤S808，判断实际货品数量与初始货品数量是否一致。如果一致，则执行步骤S812，如果不一致，则执行步骤S814。
如果实际货品数量与初始货品数量一致，则说明目标堆垛中的货品的数量没有问题，如果不需要检测其它方面，则可以正常入库；如果实际货品数量与初始货品数量不一致，则说明目标堆垛中的货品数量存在异常，可以不进行后续其它方面的检测。
如果实际货品数量与初始货品数量一致，且除了数量检测之外，还可以对目标堆垛的垛型或包装进行检测，如上述实施例三中的检测项目，这里不再赘述。此外需要说明的是，上述几种堆垛检测项目可以同时使用，也可以不同时使用，一般来说，只要有一种检测方法将目标堆垛认定为异常堆垛，即使其他的检测方法认定该目标堆垛为正常堆垛，该目标堆垛也为异常堆垛。
步骤S812，控制传送设备将目标堆垛正常入库。
步骤S814，控制传送设备将目标堆垛传送至异常处理区。
如果目标堆垛为异常堆垛，需要将其与正常堆垛放置在不同区域，因此，可以将一片区域设置为异常处理区，异常处理区负责存储异常堆垛，异常堆垛可以通过传送设备传送至异常处理区。该方式的异常堆垛不会入库，也不会和正常堆垛放置在一起，工作人员不需花费很大成本分离异常堆垛和正常堆垛。
在确定目标堆垛异常之后，服务器可以向仓库工作人员发送预警信息，提示目标堆垛异常或者 目标堆垛存入异常处理区，例如：如果目标堆垛异常，向指定的终端设备发送预警信息；其中，预警信息携带有目标堆垛的标识信息，也可以携带目标堆垛的图像(如上述多角度图像中的一个或多个)，如果目标堆垛的货品数量异常，还可以携带目标堆垛的实际货品数量。
其中，上述标识信息可以为位置标识、序号标识或者异常原因标识等，可以对异常的目标堆垛进行标识即可。工作人员在接收到预警信息之后，可以前往异常处理区对目标堆垛进行查看。
本公开实施例提供的上述方法，目标堆垛可以先由人工或其他方式(例如机械设备，RFID等)进行初始货品数量录入，目标堆垛经传送设备传送，经过入库口门架或指定监测点的门架时，门架上的摄像设备扫描目标堆垛的多角度图像，通过对目标堆垛的货品数量进行计算盘点，判断目标堆垛是否异常，该方式通过服务器可以快速、准确地进行堆垛检测，只需要很少的人力和时间成本，检测效率和检测成功率都较高，并且不需要高成本的高密度光栅设备，检测后，异常堆垛和正常堆垛传送至不同区域，工作人员不需花费很大成本分离异常堆垛和正常堆垛。
下面对入库管理的整体流程进行说明，可以参见图10所示的一种入库管理方法的整体流程示意图，如图10所示，人工输入货品信息x，服务器接收货品信息x后，可以向传送设备发送运送指令a，传送设备执行运送指令a，将堆垛运送至指定监测点。在指定监测点设置有多个摄像设备，在传送设备将堆垛运送至指定监测点之后，服务器向摄像设备发送拍照指令b，摄像设备执行拍照指令b，采集堆垛的多角度图像，发送多角度图像至服务器。服务器接收多角度图像，根据多角度图像判断堆垛是否异常。
如果堆垛不异常，服务器可以将堆垛的信息存档，向传送设备发送运送指令c，传送设备执行运送指令c，将堆垛运送入库。
如果堆垛异常，服务器可以对多角度图像中的异常区域进行标注并存档，向传送设备发送运送指令d，传送设备执行运送指令d，将堆垛运送至异常处理区，之后由工作人员进行人工处理。其中，还可以向工作人员的终端设备发送预警信息，以使工作人员在接收到预警信息后进行人工处理。
对应于上述方法实施例，参见图11所示的一种入库管理装置的结构示意图，该装置应用于服务器，也即，该装置可安装于服务器侧，该服务器与多个摄像设备通信连接；该装置包括以下模块：
第一控制模块1102，配置成在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；其中，多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛；
第二控制模块1104，配置成对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库。
本公开实施例提供的上述入库管理装置，通过在目标堆垛传送过程中，由多个摄像设备采集目标堆垛的多角度图像，对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将该目标堆垛正常入库。该方式中，服务器通过对目标堆垛的多角度图像进行图像分析，确定其是否正常入库，进而无需人工进行堆垛检测，可以节约人力成本和时间成本；通过图像分析检测多角度图像可以快速、准确地确定目标堆垛是否可以正常入库，从而提升堆垛检测的检测效率和检测准确率。
可选的，上述传送设备上设置有光电传感器，该光电传感器配置成检测目标堆垛是否到达指定监测点；基于此，上述第一控制模块1102还配置成当光电传感器监测到目标堆垛到达指定监测点时，向多个摄像设备发送图像采集指令，以使多个摄像设备采集目标堆垛的多角度图像。
可选的，上述传送设备包括直线型输送线，多个摄像设备设置在直线型输送线对应的门架上；指定监测点为门架对应的监测点。
可选的，上述多个摄像设备的数量为5个，其中，4个摄像设备分别配置成拍摄位于监测点的目标堆垛的四个侧面，1个摄像设备配置成拍摄目标堆垛的顶面。
可选的，上述传送设备包括构成L型输送线的第一输送线和第二输送线，光电传感器包括第一光电传感器和第二光电传感器，第一光电传感器和第二光电传感器分别设置在第一输送线和第二输送线上，多个摄像设备中的第一组摄像设备设置在第一输送线的第一门架上，多个摄像设备中的第二组摄像设备设置在第二输送线的第二门架上；相应地，上述第一控制模块1102还配置成：当第一光电传感器监测到目标堆垛到达第一输送线上的指定监测点时，向第一组摄像设备发送图像采集指令；当第二光电传感器监测到目标堆垛到达第二输送线上的指定监测点时，向第二组摄像设备发送图像采集指令。
可选的，上述多个摄像设备的数量为5个，其中，第一组摄像设备包括3个摄像设备，分别配置成拍摄位于第一输送线上的指定监测点的目标堆垛的两个侧面和目标堆垛的顶面；第二组摄像设备包括2个摄像设备，分别配置成拍摄位于第二输送线上的指定监测点的目标堆垛的另两个侧面。
可选的，上述第二控制模块1104还配置成：通过预先训练好的图像分析模型，对目标堆垛的多角度图像进行以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测；根据检测结果确定目标堆垛是否异常。
可选的，本公开实施例中，上述垛型检测包括以下至少之一：检测目标堆垛的倾斜程度；检测目标堆垛中的货品与托盘的摆放关系；检测目标堆垛中的货品上下摆放关系；检测目标堆垛中的货品间距。
可选的，在本公开另一实施例中，上述装置还包括：信息获取模块，与第一控制模块1102或第二控制模块1104连接，该信息获取模块配置成：获取待入库的目标堆垛的初始货品数量；相应地，第二控制模块1104还配置成：根据目标堆垛的多角度图像确定目标堆垛中包含的实际货品数量；判断实际货品数量与初始货品数量是否一致；如果不一致，确定目标堆垛异常。
可选的，上述多角度图像包括目标堆垛的4个侧面的侧面图像和1个顶面的顶面图像；相应地，第二控制模块1104还配置成：分别将每个侧面图像和顶面图像作为目标图像，在每个目标图像中标注出距离目标摄像设备最近的堆垛表面包含的货品框；其中，目标摄像设备为采集目标图像的摄像设备；基于标注后的每个目标图像包含的货品框确定目标堆垛中包含的实际货品数量。
可选的，上述第二控制模块1104还配置成：如果分析结果为目标堆垛异常，控制传送设备将目标堆垛传送至异常处理区；如果分析结果为目标堆垛正常，控制传送设备将目标堆垛正常入库。
本公开实施例提供了一种仓库管理***，包括服务器、与服务器通信连接的传送设备和多个摄像设备；该传送设备配置成在服务器的控制下传送目标堆垛；该服务器配置成：在目标堆垛在传送设备上进行传送的过程中，控制多个摄像设备采集目标堆垛的多角度图像；以及对目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将目标堆垛正常入库；多个摄像设备设置在目标堆垛的传送路径上，分别从多个不同角度拍摄目标堆垛。
可选的，上述传送设备包括直线型输送线，多个摄像设备设置在直线型输送线对应的门架上；或者，上述传送设备包括L型输送线，多个摄像设备中的第一组摄像设备设置在L型输送线对应的第一门架上，多个摄像设备中的第二组摄像设备设置在L型输送线对应的第二门架上。
本公开实施例提供了一种电子***，该电子***包括：处理设备和存储装置；存储装置上存储有计算机程序，计算机程序在被处理设备运行时执行上述的入库管理方法。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的电子***的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
本公开实施例还提供了一种计算机可读存储介质，计算机可读存储介质上存储有计算机程序，计算机程序被处理设备运行时执行如上述入库管理方法的步骤。
本公开实施例所提供的入库管理方法、装置、仓库管理***和电子***的计算机程序产品，包括存储了程序代码的计算机可读存储介质，程序代码包括的指令可用于执行前面方法实施例中的方法，具体实现可参见方法实施例，在此不再赘述。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的***和/或装置的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
另外，在本公开实施例的描述中，除非另有明确的规定和限定，术语“安装”、“相连”、“连接”应做广义理解，例如，可以是固定连接，也可以是可拆卸连接，或一体地连接；可以是机械连接，也可以是电连接；可以是直接相连，也可以通过中间媒介间接相连，可以是两个元件内部的连通。对于本领域的普通技术人员而言，可以具体情况理解上述术语在本公开中的具体含义。
功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本公开的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案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一个存储介质中，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，服务器，或者网络设备等)执行本公开各个实施例方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
在本公开的描述中，需要说明的是，术语“中心”、“上”、“下”、“左”、“右”、“竖直”、“水平”、“内”、“外”等指示的方位或位置关系为基于附图所示的方位或位置关系，仅是为了便于描述本公开和简化描述，而不是指示或暗示所指的装置或元件必须具有特定的方位、以特定的方位构造和操作，因此不能理解为对本公开的限制。此外，术语“第一”、“第二”、“第三”仅用于描述目的，而不能理解为指示或暗示相对重要性。
最后应说明的是：以上所述实施例，仅为本公开的具体实施方式，用以说明本公开的技术方案，而非对其限制，本公开的保护范围并不局限于此，尽管参照前述实施例对本公开进行了详细的说明，本领域的普通技术人员应当理解：任何熟悉本技术领域的技术人员在本公开揭露的技术范围内，其依然可以对前述实施例所记载的技术方案进行修改或可轻易想到变化，或者对其中部分技术特征进行等同替换；而这些修改、变化或者替换，并不使相应技术方案的本质脱离本公开实施例技术方案的精神和范围，都应涵盖在本公开的保护范围之内。因此，本公开的保护范围应所述以权利要求的保护范围为准。
工业实用性
本公开提出的技术方案中，服务器通过对目标堆垛的多角度图像进行图像分析，确定其是否正常入库，进而无需人工进行堆垛检测，可以节约人力成本和时间成本；通过图像分析检测多角度图像可以快速、准确地确定目标堆垛是否可以正常入库，从而提升堆垛检测的检测效率和检测准确率，保证了入库堆垛的质量。

Claims (19)

  1. 一种入库管理方法，其特征在于，所述方法应用于服务器，所述服务器与多个摄像设备通信连接；所述方法包括：
    在目标堆垛在传送设备上进行传送的过程中，控制所述多个摄像设备采集所述目标堆垛的多角度图像；所述多个摄像设备设置在所述目标堆垛的传送路径上，分别从多个不同角度拍摄所述目标堆垛；
    对所述目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将所述目标堆垛正常入库。
  2. 根据权利要求1所述的方法，其特征在于，所述传送设备上设置有光电传感器，所述光电传感器配置成检测所述目标堆垛是否到达指定监测点；
    所述控制所述多个摄像设备采集所述目标堆垛的多角度图像，包括：
    当所述光电传感器监测到所述目标堆垛到达所述指定监测点时，向所述多个摄像设备发送图像采集指令，以使所述多个摄像设备采集所述目标堆垛的多角度图像。
  3. 根据权利要求2所述的方法，其特征在于，所述传送设备包括直线型输送线，所述多个摄像设备设置在所述直线型输送线对应的门架上；所述指定监测点为所述门架对应的监测点。
  4. 根据权利要求3所述的方法，其特征在于，所述多个摄像设备的数量为5个，其中，4个摄像设备分别配置成拍摄位于所述监测点的所述目标堆垛的四个侧面，1个摄像设备配置成拍摄所述目标堆垛的顶面。
  5. 根据权利要求2所述的方法，其特征在于，所述传送设备包括构成L型输送线的第一输送线和第二输送线，所述光电传感器包括第一光电传感器和第二光电传感器，所述第一光电传感器和所述第二光电传感器分别设置在所述第一输送线和所述第二输送线上，所述多个摄像设备中的第一组摄像设备设置在所述第一输送线的第一门架上，所述多个摄像设备中的第二组摄像设备设置在所述第二输送线的第二门架上；
    当所述光电传感器监测到所述目标堆垛到达所述指定监测点时，向所述多个摄像设备发送图像采集指令，包括：
    当所述第一光电传感器监测到所述目标堆垛到达所述第一输送线上的指定监测点时，向所述第一组摄像设备发送图像采集指令；
    当所述第二光电传感器监测到所述目标堆垛到达所述第二输送线上的指定监测点时，向所述第二组摄像设备发送图像采集指令。
  6. 根据权利要求5所述的方法，其特征在于，所述多个摄像设备的数量为5个，其中，第一组摄像设备包括3个摄像设备，分别配置成拍摄位于所述第一输送线上的指定监测点的所述目标堆垛的两个侧面和所述目标堆垛的顶面；
    所述第二组摄像设备包括2个摄像设备，分别配置成拍摄位于所述第二输送线上的指定监测点的所述目标堆垛的另两个侧面。
  7. 根据权利要求1至6任一项所述的方法，其特征在于，对所述目标堆垛的多角度图像进行图像分析，包括：
    通过预先训练好的图像分析模型，对所述目标堆垛的多角度图像进行以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测；
    根据检测结果确定所述目标堆垛是否异常。
  8. 根据权利要求1至7任一项所述的方法，其特征在于，所述图像分析包括垛型检测，所述垛型检测包括以下至少之一：
    检测所述目标堆垛的倾斜程度；
    检测所述目标堆垛中的货品与托盘的摆放关系；
    检测所述目标堆垛中的货品上下摆放关系；
    检测所述目标堆垛中的货品间距。
  9. 根据权利要求1至8任一项所述的方法，其特征在于，所述对所述目标堆垛的多角度图像进行图像分析之前，所述方法还包括：获取待入库的目标堆垛的初始货品数量；
    所述对所述目标堆垛的多角度图像进行图像分析，包括：
    根据所述目标堆垛的多角度图像确定所述目标堆垛中包含的实际货品数量；
    判断所述实际货品数量与所述初始货品数量是否一致；
    如果不一致，确定所述目标堆垛异常。
  10. 根据权利要求9所述的方法，其特征在于，所述多角度图像包括所述目标堆垛的4个侧面的侧面图像和1个顶面的顶面图像；
    根据所述目标堆垛的多角度图像确定所述目标堆垛中包含的实际货品数量的步骤，包括：
    分别将每个所述侧面图像和所述顶面图像作为目标图像，在每个所述目标图像中标注出距离目标摄像设备最近的堆垛表面包含的货品框；其中，所述目标摄像设备为采集所述目标图像的摄像设备；
    基于标注后的每个所述目标图像包含的货品框确定所述目标堆垛中包含的实际货品数量。
  11. 根据权利要求1-10任一项所述的方法，其特征在于，所述根据分析结果确定是否将所述目标堆垛正常入库，包括：
    如果分析结果为所述目标堆垛异常，控制所述传送设备将所述目标堆垛传送至异常处理区；如果分析结果为所述目标堆垛正常，控制所述传送设备将所述目标堆垛正常入库。
  12. 一种入库管理装置，其特征在于，所述装置应用于服务器，所述服务器与多个摄像设备通信连接；所述装置包括：
    第一控制模块，配置成在目标堆垛在传送设备上进行传送的过程中，控制所述多个摄像设备采集所述目标堆垛的多角度图像；所述多个摄像设备设置在所述目标堆垛的传送路径上，分别从多个不同角度拍摄所述目标堆垛；
    第二控制模块，配置成对所述目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将所述目标堆垛正常入库。
  13. 根据权利要求12所述的装置，其特征在于，所述第二控制模块还配置成：
    通过预先训练好的图像分析模型，对所述目标堆垛的多角度图像进行以下至少一种检测：货品数量检测、垛型检测、货品包装变形检测、货品包装破损检测和货品包装污迹检测；根据检测结果确定所述目标堆垛是否异常。
  14. 根据权利要求12或13所述的装置，其特征在于，所述装置还包括：信息获取模块，配置成获取待入库的目标堆垛的初始货品数量；
    所述第二控制模块还配置成：根据所述目标堆垛的多角度图像确定所述目标堆垛中包含的实际货品数量；判断所述实际货品数量与所述初始货品数量是否一致；如果不一致，确定所述目标堆垛异常。
  15. 根据权利要求12至14任一项所述的装置，其特征在于，所述第二控制模块还配置成：
    如果分析结果为所述目标堆垛异常，控制所述传送设备将所述目标堆垛传送至异常处理区；如果分析结果为所述目标堆垛正常，控制所述传送设备将所述目标堆垛正常入库。
  16. 一种仓库管理***，其特征在于，包括服务器、与所述服务器通信连接的传送设备和多个摄像设备；
    所述传送设备配置成在所述服务器的控制下传送目标堆垛；
    所述服务器配置成：在所述目标堆垛在所述传送设备上进行传送的过程中，控制所述多个摄像设备采集所述目标堆垛的多角度图像；以及对所述目标堆垛的多角度图像进行图像分析，根据分析结果确定是否将所述目标堆垛正常入库；
    所述多个摄像设备设置在所述目标堆垛的传送路径上，分别从多个不同角度拍摄所述目标堆垛。
  17. 根据权利要求16所述的***，其特征在于，所述传送设备包括直线型输送线，所述多个摄像设备设置在所述直线型输送线对应的门架上；或者，
    所述传送设备包括L型输送线，所述多个摄像设备中的第一组摄像设备设置在所述L型输送线对应的第一门架上，所述多个摄像设备中的第二组摄像设备设置在所述L型输送线对应的第二门架上。
  18. 一种电子***，其特征在于，所述电子***包括：处理设备和存储装置；
    所述存储装置上存储有计算机程序，所述计算机程序在被所述处理设备运行时执行如权利要求1至11任一项所述的入库管理方法。
  19. 一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，其特征在于，所述计算机程序被处理设备运行时执行如权利要求1至11任一项所述的入库管理方法的步骤。
PCT/CN2021/102425 2020-11-23 2021-06-25 入库管理方法、装置、仓库管理***和电子*** WO2022105231A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011325601.1A CN112651297A (zh) 2020-11-23 2020-11-23 入库管理方法、装置、仓库管理***和电子***
CN202011325601.1 2020-11-23

Publications (1)

Publication Number Publication Date
WO2022105231A1 true WO2022105231A1 (zh) 2022-05-27

Family

ID=75349308

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/102425 WO2022105231A1 (zh) 2020-11-23 2021-06-25 入库管理方法、装置、仓库管理***和电子***

Country Status (2)

Country Link
CN (1) CN112651297A (zh)
WO (1) WO2022105231A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115511875A (zh) * 2022-10-28 2022-12-23 上海东普信息科技有限公司 货物堆积的检测方法、装置、设备及存储介质
CN116308070A (zh) * 2023-05-24 2023-06-23 福建省万物智联科技有限公司 一种次品收集管理的方法及相关设备

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112651297A (zh) * 2020-11-23 2021-04-13 北京旷视机器人技术有限公司 入库管理方法、装置、仓库管理***和电子***
CN113240892A (zh) * 2021-04-26 2021-08-10 深圳市安星数字***有限公司 夜视码头监测方法、装置、设备及存储介质
CN114044296A (zh) * 2021-12-20 2022-02-15 深圳市海柔创新科技有限公司 仓储***及货箱入库方法
CN114078127B (zh) * 2022-01-20 2022-05-17 广州里工实业有限公司 物件的缺陷检测及计数方法、装置、设备和存储介质
CN115049620B (zh) * 2022-06-17 2024-03-26 中轻长泰(长沙)智能科技股份有限公司 一种纸垛外观检测方法、装置、设备及存储介质

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106210616A (zh) * 2015-05-04 2016-12-07 杭州海康威视数字技术股份有限公司 集装箱图像信息的采集方法、装置和***
CN106695792A (zh) * 2017-01-05 2017-05-24 中国计量大学 基于机器视觉的码垛机器人跟踪监控***及方法
WO2018022265A1 (en) * 2016-07-28 2018-02-01 X Development Llc Inventory management
CN109344828A (zh) * 2018-09-30 2019-02-15 中铁武汉勘察设计研究院有限公司 一种集装箱图像采集***及集装箱装载状态检测方法
CN110065754A (zh) * 2019-04-30 2019-07-30 南京斯杩克机器人技术有限公司 一种大宗物料堆垛视觉扫描***及其图像处理方法
CN111625001A (zh) * 2020-05-28 2020-09-04 珠海格力智能装备有限公司 机器人的控制方法及装置、工业机器人
CN112651297A (zh) * 2020-11-23 2021-04-13 北京旷视机器人技术有限公司 入库管理方法、装置、仓库管理***和电子***

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109870983B (zh) * 2017-12-04 2022-01-04 北京京东尚科信息技术有限公司 处理托盘堆垛图像的方法、装置及用于仓储拣货的***
CN110977939B (zh) * 2019-11-26 2021-05-04 重庆凡聚智能科技有限公司 目标工件识别定位***

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106210616A (zh) * 2015-05-04 2016-12-07 杭州海康威视数字技术股份有限公司 集装箱图像信息的采集方法、装置和***
WO2018022265A1 (en) * 2016-07-28 2018-02-01 X Development Llc Inventory management
CN106695792A (zh) * 2017-01-05 2017-05-24 中国计量大学 基于机器视觉的码垛机器人跟踪监控***及方法
CN109344828A (zh) * 2018-09-30 2019-02-15 中铁武汉勘察设计研究院有限公司 一种集装箱图像采集***及集装箱装载状态检测方法
CN110065754A (zh) * 2019-04-30 2019-07-30 南京斯杩克机器人技术有限公司 一种大宗物料堆垛视觉扫描***及其图像处理方法
CN111625001A (zh) * 2020-05-28 2020-09-04 珠海格力智能装备有限公司 机器人的控制方法及装置、工业机器人
CN112651297A (zh) * 2020-11-23 2021-04-13 北京旷视机器人技术有限公司 入库管理方法、装置、仓库管理***和电子***

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115511875A (zh) * 2022-10-28 2022-12-23 上海东普信息科技有限公司 货物堆积的检测方法、装置、设备及存储介质
CN116308070A (zh) * 2023-05-24 2023-06-23 福建省万物智联科技有限公司 一种次品收集管理的方法及相关设备
CN116308070B (zh) * 2023-05-24 2023-09-05 福建省万物智联科技有限公司 一种次品收集管理的方法及相关设备

Also Published As

Publication number Publication date
CN112651297A (zh) 2021-04-13

Similar Documents

Publication Publication Date Title
WO2022105231A1 (zh) 入库管理方法、装置、仓库管理***和电子***
US10984378B1 (en) Profiling pallets and goods in a warehouse environment
CN106573381B (zh) 卡车卸载机可视化
CN111861306A (zh) 具有自动化包裹登记机构和自动检测流水线的机器人***
WO2021218792A1 (zh) 包裹处理设备、包裹处理方法、电子设备及存储介质
WO2022206744A1 (zh) 信息关联方法、***、装置、服务器及存储介质
CN102232222A (zh) 用于接收装运包裹的***和方法
CN112507820A (zh) 自动盘点货物的方法、装置、***和电子设备
JP2017171444A (ja) 投影指示装置、荷物仕分けシステムおよび投影指示方法
CN110888345A (zh) 检测设备、控制设备及自动校验***和方法
CN112099104A (zh) 一种安检方法及装置
JP2012137304A (ja) 物流システムにおける自動計測装置
CN111274951A (zh) 料箱状态的监控方法及装置、自动投料***
US20240166450A1 (en) Cased goods inspection system and method therefor
US20240139778A1 (en) Methods, apparatuses, and systems for automatically performing sorting operations
CN114529843A (zh) 货物拥堵识别方法、装置、电子设备和存储介质
CN112184104A (zh) 一种用于仓储的物料码放方法
JP2021025853A (ja) サイズ測定装置、サイズ測定方法、プログラム
JP7288568B1 (ja) 自動採寸システム
CN210585910U (zh) 包裹粗分设备及自动分拣***
JP7366195B2 (ja) ケージ積上げの確認方法、コンピュータ装置及び記憶媒体
WO2022145236A1 (ja) 情報処理装置及びプログラム
KR101150754B1 (ko) 영상처리를 이용한 약품상자 영역 검출 시스템 및 방법
CN113978987A (zh) 栈板物品打包提货方法、装置、设备及介质
KR20230069267A (ko) 택배화물 취급 자동화를 위한 화물정보 인식 장치 및 방법

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: 21893384

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21893384

Country of ref document: EP

Kind code of ref document: A1