CN108596221A - The image-recognizing method and equipment of rod reading - Google Patents
The image-recognizing method and equipment of rod reading Download PDFInfo
- Publication number
- CN108596221A CN108596221A CN201810315600.5A CN201810315600A CN108596221A CN 108596221 A CN108596221 A CN 108596221A CN 201810315600 A CN201810315600 A CN 201810315600A CN 108596221 A CN108596221 A CN 108596221A
- Authority
- CN
- China
- Prior art keywords
- scale
- image
- length
- measured
- images
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000001514 detection method Methods 0.000 claims abstract description 44
- 238000005259 measurement Methods 0.000 claims abstract description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 18
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 10
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 241000406668 Loxodonta cyclotis Species 0.000 claims 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 22
- 238000010586 diagram Methods 0.000 description 14
- 238000013135 deep learning Methods 0.000 description 6
- 238000012545 processing Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 4
- 238000000605 extraction Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000001629 suppression Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 241001269238 Data Species 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000008309 brain mechanism Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000877 morphologic effect Effects 0.000 description 1
- 230000008450 motivation Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The present invention provides a kind of image-recognizing method of rod reading and equipment.This method includes：Obtain the image using tape measure object to be measured；Length of the scale image in the image for measuring object to be measured along scale measurement direction is identified using scale image detection model, and intercepts the scale image from the image for measuring object to be measured；The length in direction is measured along the scale using the scale images in the scale image of scale images detection model identification interception；The physical length that the length in direction, the scale images measure along scale scale along the length, the actual total length of the scale and the scale in direction is measured along scale according to the scale image, the reading of object to be measured described in the tape measure is calculated.By identifying the survey measurements of the length of scale and the length computation scale of scale, adaptability of the scale image recognition to complex environment can be improved.
Description
Technical field
The present invention relates to image identification technical field more particularly to the image-recognizing methods and equipment of a kind of rod reading.
Background technology
With the development of information age, image data is more huge, and is being increased with a speed quickly.Pass through this
A little images can excavate out the thinking of people, and foundation is provided for further decision.Before carrying out data mining, need pair
These huge image datas carry out effective Classification Management.In the Classification Management to image data, image recognition is one
Element task has important meaning and value in terms of information extraction, information identification and information retrieval.Allow computer picture
People equally carries out image classification or image recognition to image data, and there is prodigious difficulty, these difficulties to be that image data has
Scrambling, the representation method of different images data, order of magnitude difference of image etc..Deep learning (Deep
Learning), especially convolutional neural networks (CNN), are the research emphasis in Image Processing and Pattern Recognition in recent years, by
It is more and more paid close attention to people.It is a kind of neural network established, simulate human brain progress analytic learning, can imitate human brain
Mechanism explain data, such as image, sound and text.
Scale image intelligent identifying system is a towards water industry and energy industry in numerous image recognition products
Scale label image automatic identification product, reading that can be in automatic identification river gage picture, and the achievement of identification carried out special
The service application of industry can substitute traditional artificial observation and count, improve user job efficiency, reduce manual operation.
From the point of view of the hierarchical structure of machine learning model, the development of machine learning probably experienced to be changed twice：Shallow-layer
Habit and deep learning.When in face of foreign lands' problem of multiple variables, shallow-layer learning model is difficult to express it, needs multilayer
The network of implicit node could be effectively indicated, and deep learning model succinctly can effectively indicate complicated function.With
The development of big data, huge data are faced, shallow-layer learning structure is even more to have shown short slab in terms of model descriptive power,
It is difficult to the intrinsic representation of abundant mining data, only the stronger model of ability to express could be excavated out from big data more more has
The information of value, this also excites the learning motivation that people explore deep learning model to Complex Function Modeling.Image recognition
Goal in research is exactly to be divided into pre-defined different classifications according to certain possessed attribute in image.How to build
Image feature representation and disaggregated model are the key that solve the problems, such as image understanding, and Many researchers are conducted extensive research and carried
Some effective methods are gone out.Traditional method is largely view-based access control model code book model, which is utilized manually well
The iamge description of ingehious design and effective machine learning model, but its expression to image media layer damage and high-layer semantic information
Power is limited, can not break through semantic gap.In recent years, the breakthrough development of deep learning provides new think of to solve this problem
Road, and application of succeeding in many pattern recognition problems.
However, the existing image recognition product based on deep learning, the powerful hypothesis space of network model makes mould
Type training is easily absorbed in local optimum, leads to model generalization poor ability, can not predict unknown data well.Meanwhile
Largely the data containing semantic label are trained model training needs, increase artificial mark workload.It is existing to be based on
Hough converts the scale recognition methods with harris detections, after algorithm is using medium filtering removal noise and gray scale balance,
It reuses morphologic refinement and outline extraction technique calculates the position of groove.The algorithm is using traditional image processing method
Method only does very well on a small amount of image, can not well adapt to various water gauge models (for example, water gauge color, scale, size
Deng) and complex environment (for example, illumination, angle etc.) variation.The existing scale location algorithm based on scale label color, is adopted
Row threshold division is gone forward side by side to be positioned to scale with Color Channel feature in the method for color segmentation extraction image, but the algorithm
Complex environment cannot be well adapted to, for example, there are stains, spot, local damage or scale color are various above scale
When, the robustness of algorithm is not high.
Invention content
The present invention provides a kind of image-recognizing method of rod reading and equipment, to improve scale image recognition to complexity
The adaptability of environment.
An embodiment of the present invention provides a kind of image-recognizing methods of rod reading, including：Acquisition is waited for using tape measure
Survey the image of object；Identify the scale image in the image for measuring object to be measured along scale using scale image detection model
The length in direction is measured, and the scale image is intercepted from the image for measuring object to be measured；It is detected using scale images
Scale images in the scale image of Model Identification interception measure the length in direction along the scale；According to scale image edge
Scale measures the length in direction, the scale images measure length, the actual total length of the scale and the institute in direction along scale
The reading of object to be measured described in the tape measure is calculated in the physical length for stating scale in scale.
In one embodiment, this method further includes：Using the first training data and image object detection algorithm, to initial god
It is trained through network, obtains the scale image detection model；First training data includes tape measure image and phase
The scale box label answered.
In one embodiment, this method further includes：Using the second training data and image object detection algorithm, to initial god
It is trained through network, obtains the scale images detection model；Second training data includes in tape measure image
Scale images and corresponding scale box label.
In one embodiment, using the scale images in the scale image of scale images detection model identification interception described in
Scale measures the length in direction, including：At least one of scale image using the identification interception of scale images detection model mark
Spend the scale frame coordinate of image；Each scale images are calculated along the scale according to each scale frame coordinate
Measure the initial length in direction；The scale of the interception is obtained according to the corresponding initial length of at least one scale images
Scale images in image measure the length in direction along the scale.
In one embodiment, the initial neural network is convolutional neural networks；Described image algorithm of target detection is
Faster RCNN algorithm of target detection；The method further includes：Based on setting network structure, obtained using ZFNet algorithms described
Initial neural network.
In one embodiment, the reading of object to be measured described in the tape measure is：
Hw=H-Hr/He*scale,
Wherein, Hw indicates that the reading of object to be measured described in the tape measure, H indicate the actual total length of the scale,
Hr indicates that the scale image measures the length in direction along scale, and He indicates that the scale images measure the length in direction along scale
Degree, scale indicate the physical length of scale in the scale.
In one embodiment, this method further includes：When the reading is more than setting value, sends warning information to terminal and set
It is standby.
In one embodiment, the image using tape measure object to be measured is obtained, including：It obtains to be measured using tape measure
The video image of object；The video image is converted into the image for measuring object to be measured.
The embodiment of the present invention also provides a kind of computer readable storage medium, is stored thereon with computer program, the program
The step of the various embodiments described above the method is realized when being executed by processor.
The embodiment of the present invention also provides a kind of computer equipment, including memory, processor and storage are on a memory simultaneously
The computer program that can be run on a processor, the processor realize the various embodiments described above the method when executing described program
The step of.
Image-recognizing method, computer readable storage medium and the computer equipment of the rod reading of the embodiment of the present invention,
It, can not be by subject surface to be measured light by identifying the survey measurements of the length of scale and the length computation scale of scale
The influence of variation, measurement angle etc. improves adaptability of the scale image recognition to complex environment.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
Obtain other attached drawings according to these attached drawings.In the accompanying drawings：
Fig. 1 is the flow diagram of the image-recognizing method of the rod reading of one embodiment of the invention.
Fig. 2 is that the scale figure in the scale image of scale images detection model identification interception is utilized in one embodiment of the invention
As the method flow schematic diagram for the length for measuring direction along scale.
Fig. 3 is the flow diagram of the image-recognizing method of the rod reading of another embodiment of the present invention.
Fig. 4 is the method flow schematic diagram that the image using tape measure object to be measured is obtained in one embodiment of the invention.
Specific implementation mode
Understand in order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the accompanying drawings to this hair
Bright embodiment is described in further details.Here, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but simultaneously
It is not as a limitation of the invention.
Fig. 1 is the flow diagram of the image-recognizing method of the rod reading of one embodiment of the invention.As shown in Figure 1, should
The image-recognizing method of rod reading, it may include：
Step S110：Obtain the image using tape measure object to be measured；
Step S120：The scale image edge in the image for measuring object to be measured is identified using scale image detection model
Scale measures the length in direction, and intercepts the scale image from the image for measuring object to be measured；
Step S130：Using the scale images in the scale image of scale images detection model identification interception along the scale
Measure the length in direction；
Step S140：The length in direction is measured along scale, the scale images are measured along scale according to the scale image
The physical length of scale in the length in direction, the actual total length of the scale and the scale is calculated the scale and surveys
Measure the reading of the object to be measured.
The scale is, for example, water gauge, which is, for example, reservoir, river etc..The scale can be on cuing scale a degree
Measure unit, the scale E for example, on water gauge.When object to be measured using tape measure, generally only part scale is it is seen, for example, profit
When measuring reservoir depth with water gauge, the water gauge of general only water surface above section as it can be seen that so, the figure for measuring object to be measured
Normally only include the scale of visible part in scale image as in.The scale image measures length and the institute in direction along scale
Length of the scale images along scale measurement direction is stated to be determined according to the distance of point-to-point transmission in image.The reality of the scale is total
The physical length of scale can be determined according to true scale in length and the scale.In the image for measuring object to be measured
Can include one or more scale images.Scale image gets the bid ruler part should be comprising one or more clearly scales, scale
In image in scale portion can there are one or multiple scales it is damaged or missing.The scale image detection model and the scale images
Detection model can be the parameter model for certain network structure.
In the present embodiment, the scale in the image for measuring object to be measured is identified by using scale image detection model
Image measures the length in direction along scale, and intercepts the scale image from the image for measuring object to be measured, utilizes mark
The scale images spent in the scale image of image detection Model Identification interception measure the length in direction along the scale, and according to institute
It states scale image and measures length, the reality of the scale that the length in direction, the scale images measure along scale direction along scale
The physical length of scale calculates the reading of tape measure object to be measured in border total length and the scale, realizes and passes through knowledge
Not Chu Lai scale length and scale length computation scale survey measurements, not by subject surface to be measured light variation, measure
The influence of angle etc. improves adaptability of the scale image recognition to complex environment.
In embodiment, in above-mentioned steps S120 and step S130, when prediction, takes the confidence threshold value conf_thresh can be
0.7~0.9, such as 0.80 is taken, non-maxima suppression threshold value nms_thresh can be 0.2~0.4, such as take 0.30.
In some embodiments, the image-recognizing method of rod reading may also include：Utilize the first training data and image mesh
Detection algorithm is marked, initial neural network is trained, the scale image detection model is obtained；The first training data packet
Include tape measure image and corresponding scale box label.Tape measure image refers to the image using tape measure object to be measured.
Each scale box label can indicate the position of scale in tape measure image, can outline scale and come.
In some embodiments, the image-recognizing method of rod reading may also include：Utilize the second training data and image mesh
Detection algorithm is marked, initial neural network is trained, the scale images detection model is obtained；The second training data packet
Include the scale images in tape measure image and corresponding scale box label.Each scale box label can indicate scale images
Position or frame coordinate.
Above-mentioned second training data can be obtained based on above-mentioned first training data.Obtain the scale image detection model
It can be identical with the initial neural network for obtaining used in the scale images detection model.When training, iterations, such as can be with
It is set as 50,000 times.
Fig. 2 is that the scale figure in the scale image of scale images detection model identification interception is utilized in one embodiment of the invention
As the method flow schematic diagram for the length for measuring direction along scale.As shown in Fig. 2, in above-mentioned steps S130, scale figure is utilized
Identify that the scale images in the scale image of interception measure the length in direction along the scale as detection model, it may include：
Step S131：At least one of scale image using the identification interception of scale images detection model scale images
Frame coordinate；
Step S132：Each scale images are calculated according to each scale frame coordinate to measure along the scale
The initial length in direction；
Step S133：The scale of the interception is obtained according to the corresponding initial length of at least one scale images
Scale images in image measure the length in direction along the scale.
Scale portion in the scale image of the interception can include one or more clearly scales.The scale frame coordinate
It may include the coordinate put on two or more scale frames.The scale images measure the initial length in direction along the scale
It can be the difference of 2 ordinates in the scale frame coordinate.In above-mentioned steps S133, for example, can be by multiple institutes
The corresponding initial length of scale images is stated to be averaging to obtain the length that scale images measure direction along the scale；Or choosing
Take length of one of initial length as scale images along scale measurement direction.
In some embodiments, the initial neural network is convolutional neural networks.In some embodiments, described image target
Detection algorithm is Faster RCNN algorithm of target detection.
In some embodiments, the image-recognizing method of rod reading may also include：Based on setting network structure, use
ZFNet algorithms obtain the initial neural network.
In some embodiments, the reading of object to be measured described in the tape measure is：
Hw=H-Hr/He*scale,
Wherein, Hw indicates that the reading of object to be measured described in the tape measure, H indicate the actual total length of the scale,
Hr indicates that the scale image measures the length in direction along scale, and He indicates that the scale images measure the length in direction along scale
Degree, scale indicate the physical length of scale in the scale.
In the embodiment, for example, when measuring water depth using water gauge, the water surface or more is subtracted using the total length of water gauge
The length of part water gauge (Hr/He*scale) can obtain the reading that water gauge measures water body, the i.e. depth of water.
Fig. 3 is the flow diagram of the image-recognizing method of the rod reading of another embodiment of the present invention.As shown in figure 3,
The image-recognizing method of rod reading shown in FIG. 1, may also include：
Step S150：When the reading is more than setting value, warning information is sent to terminal device.
The terminal device can be mobile terminal device, such as mobile phone.Further, it is possible to which warning information is sent to terminal
In APP (application) in equipment, such as short message, wechat etc..With this, object pair to be measured can be being learnt by tape measure reading
When people generate threat, related personnel is prompted to make counter-measure in time.
Fig. 4 is the method flow schematic diagram that the image using tape measure object to be measured is obtained in one embodiment of the invention.
As shown in figure 4, in above-mentioned steps S110, the image using tape measure object to be measured is obtained, it may include：
Step S111：Obtain the video image using tape measure object to be measured；
Step S112：The video image is converted into the image for measuring object to be measured.
In the embodiment, video image can be obtained from the monitoring point of object to be measured.It can be by being cut from video image
Figure obtains the image for measuring object to be measured.It is more by that can be obtained in the video pictures progress sectional drawing of different play times
A image for measuring object to be measured.
In one specific embodiment, by taking water gauge as an example, the water gauge image automatic identification algorithm based on deep learning can wrap
It includes：
Step S1：Water gauge image recognition network structure is defined, algorithm uses training network based on ZFNet；
Step S2：According to the network structure and parameter of step S1 configurations, using faster rcnn algorithm of target detection to mark
Ruler is trained respectively with scale E, and iterations are disposed as 50000 times, obtains scale and scale E detection models；
Step S3：The scale detection model obtained according to step S2 is sent into network with input picture I and carries out forward calculation, in advance
It is 0.80 that confidence threshold value conf_thresh is taken when survey, and non-maxima suppression threshold value nms_thresh is 0.30, obtains scale
It returns frame coordinate and intercepts out scale image；
Step S4：The scale E detection models obtained according to step S2 and the scale image of step S3 interceptions are sent into convolution god
Forward calculation is carried out through network, it is 0.80 that when prediction, which takes confidence threshold value conf_thresh, non-maxima suppression threshold value nms_
Thresh is 0.30, obtains the recurrence frame coordinate BBoxi (x1, y1, x2, y2) of N number of scale E on scale, and it is average to calculate scale E frames
Height calculation formula is He=1/N* ∑s (y2-y1)；
Step S5：The frame height of the position and scale image acceptance of the bid ruler of scale is obtained according to step S3 and step S4 forward calculations
Hr and scale E frame average height He are spent, the unit length to drift slide E is scale, the total a height of H of water gauge, then depth of water calculation formula
For：Hw=H-Hr/He*scale (unit m).
The method of the present embodiment utilizes all kinds of river gages in computer picture automatic identification technology automatic identification water industry
Reading to replace traditional artificial naked eyes to judge reading, and sets early-warning conditions, automatically analyzes early warning, can gradually reduce river gage
The artificial energy input of observation, really promotes working efficiency.River gage video monitoring ability and dynamics can be substantially improved, find
Exception can much sooner, accurately.The effect at river gage video monitoring station can be played utmostly.
In some embodiments, the scale image intelligent identifying system of the image-recognizing method based on rod reading can mainly wrap
Include the access part of video image, image recognition computation model, result of calculation analysis and early warning part, result of calculation exposition.
The access of video image can refer to that scale image intelligent identifying system is connect with video monitoring point device, obtain video figure in real time
As information, system automatically preserves video image conversion picture, is identified for image content.Image recognition computation model can be
Automatic identification computation model is established for river gage image, allows computer that can automatically identify the reading of river gage in picture,
And constantly self-teaching optimization, improve recognition accuracy.Result of calculation analysis and early warning can be directed to the rod reading identified,
By setting threshold value of warning, automatic decision whether superthreshold, can automatically generate warning information after superthreshold, and by wechat,
Short message sending related personnel.Result of calculation displaying can be the achievement information calculated for system automatic identification, be carried out by web
Intuitive displaying and inquiry.
The embodiment of the present invention also provides a kind of computer readable storage medium, is stored thereon with computer program, the program
The step of the various embodiments described above the method is realized when being executed by processor.
The embodiment of the present invention also provides a kind of computer equipment, including memory, processor and storage are on a memory simultaneously
The computer program that can be run on a processor, the processor realize the various embodiments described above the method when executing described program
The step of.
In conclusion the image-recognizing method of the rod reading of the embodiment of the present invention, computer readable storage medium and meter
Calculating machine equipment can not be by be measured right by identifying the survey measurements of the length of scale and the length computation scale of scale
As the influence of the variation of surface light, measurement angle etc., adaptability of the scale image recognition to complex environment is improved.
In the description of this specification, reference term " one embodiment ", " specific embodiment ", " some implementations
Example ", " such as ", the description of " example ", " specific example " or " some examples " etc. mean it is described in conjunction with this embodiment or example
Particular features, structures, materials, or characteristics are included at least one embodiment or example of the invention.In the present specification,
Schematic expression of the above terms may not refer to the same embodiment or example.Moreover, the specific features of description, knot
Structure, material or feature can be combined in any suitable manner in any one or more of the embodiments or examples.Each embodiment
Involved in the step of implementation of the sequence for schematically illustrating the present invention, sequence of steps therein is not construed as limiting, can be as needed
It appropriately adjusts.
It should be understood by those skilled in the art that, the embodiment of the present invention can be provided as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, the present invention can be used in one or more wherein include computer usable program code computer
The computer program production implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of product.
The present invention be with reference to according to the method for the embodiment of the present invention, the flow of equipment (system) and computer program product
Figure and/or block diagram describe.It should be understood that can be realized by computer program instructions every first-class in flowchart and/or the block diagram
The combination of flow and/or box in journey and/or box and flowchart and/or the block diagram.These computer programs can be provided
Instruct the processor of all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine so that the instruction executed by computer or the processor of other programmable data processing devices is generated for real
The device for the function of being specified in present one flow of flow chart or one box of multiple flows and/or block diagram or multiple boxes.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that instruction generation stored in the computer readable memory includes referring to
Enable the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one box of block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device so that count
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, in computer or
The instruction executed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Particular embodiments described above has carried out further in detail the purpose of the present invention, technical solution and advantageous effect
Describe in detail it is bright, it should be understood that the above is only a specific embodiment of the present invention, the guarantor being not intended to limit the present invention
Range is protected, all within the spirits and principles of the present invention, any modification, equivalent substitution, improvement and etc. done should be included in this
Within the protection domain of invention.
Claims (10)
1. a kind of image-recognizing method of rod reading, which is characterized in that including：
Obtain the image using tape measure object to be measured；
Identify that the scale image in the image for measuring object to be measured measures direction along scale using scale image detection model
Length, and intercept the scale image from the image for measuring object to be measured；
The length in direction is measured along the scale using the scale images in the scale image of scale images detection model identification interception
Degree；
According to the scale image length, the institute that the length in direction, the scale images measure along scale direction are measured along scale
The actual total length of scale and the physical length of scale in the scale are stated, it is to be measured right described in the tape measure to be calculated
The reading of elephant.
2. the image-recognizing method of rod reading as described in claim 1, which is characterized in that further include：
Using the first training data and image object detection algorithm, initial neural network is trained, obtains the scale map
As detection model；First training data includes tape measure image and corresponding scale box label.
3. the image-recognizing method of rod reading as described in claim 1, which is characterized in that further include：
Using the second training data and image object detection algorithm, initial neural network is trained, obtains the scale figure
As detection model；Second training data includes scale images and corresponding scale box label in tape measure image.
4. the image-recognizing method of rod reading as described in claim 1, which is characterized in that utilize scale images detection model
Identify that the scale images in the scale image of interception measure the length in direction along the scale, including：
The scale frame coordinate of at least one of scale image using the identification interception of scale images detection model scale images；
Initial length of each scale images along scale measurement direction is calculated according to each scale frame coordinate
Degree；
The scale in the scale image of the interception is obtained according to the corresponding initial length of at least one scale images
Image measures the length in direction along the scale.
5. the image-recognizing method of rod reading as claimed in claim 2 or claim 3, which is characterized in that the initial neural network
For convolutional neural networks；Described image algorithm of target detection is Faster RCNN algorithm of target detection；
The method further includes：
Based on setting network structure, the initial neural network is obtained using ZFNet algorithms.
6. the image-recognizing method of rod reading as described in claim 1, which is characterized in that be measured described in the tape measure
The reading of object is：
Hw=H-Hr/He*scale,
Wherein, Hw indicates that the reading of object to be measured described in the tape measure, H indicate the actual total length of the scale, Hr tables
Showing that the scale image measures the length in direction along scale, He indicates that the scale images measure the length in direction along scale,
Scale indicates the physical length of scale in the scale.
7. the image-recognizing method of rod reading as described in claim 1, which is characterized in that further include：
When the reading is more than setting value, warning information is sent to terminal device.
8. the image-recognizing method of rod reading as described in claim 1, which is characterized in that obtain to be measured using tape measure
The image of object, including：
Obtain the video image using tape measure object to be measured；
The video image is converted into the image for measuring object to be measured.
9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor
The step of claim 1 to 8 the method is realized when row.
10. a kind of computer equipment, including memory, processor and storage are on a memory and the meter that can run on a processor
Calculation machine program, which is characterized in that the step of processor realizes claim 1 to 8 the method when executing described program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810315600.5A CN108596221B (en) | 2018-04-10 | 2018-04-10 | Image recognition method and device for scale reading |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810315600.5A CN108596221B (en) | 2018-04-10 | 2018-04-10 | Image recognition method and device for scale reading |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108596221A true CN108596221A (en) | 2018-09-28 |
| CN108596221B CN108596221B (en) | 2020-12-01 |
Family
ID=63621470
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810315600.5A Active CN108596221B (en) | 2018-04-10 | 2018-04-10 | Image recognition method and device for scale reading |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108596221B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214806A (en) * | 2018-11-20 | 2019-01-15 | 北京京东尚科信息技术有限公司 | Self-help settlement method, apparatus and storage medium |
| CN109784259A (en) * | 2019-01-08 | 2019-05-21 | 江河瑞通（北京）技术有限公司 | Water transparency intelligent identification Method and Sai Shi disk component based on image recognition |
| CN109859183A (en) * | 2019-01-29 | 2019-06-07 | 江河瑞通（北京）技术有限公司 | More element integrated water body intelligent identification Methods and ecology station based on edge calculations |
| CN110223341A (en) * | 2019-06-14 | 2019-09-10 | 北京国信华源科技有限公司 | A kind of Intelligent water level monitoring method based on image recognition |
| CN111877105A (en) * | 2020-06-16 | 2020-11-03 | 中交第三公路工程局有限公司 | Modified asphalt construction method |
| CN113112514A (en) * | 2021-04-27 | 2021-07-13 | 汇鸿智能科技(辽宁)有限公司 | Method and device for AI (Artificial Intelligence) recognition of graphite size, computer equipment and storage medium |
| CN113776408A (en) * | 2021-09-13 | 2021-12-10 | 北京邮电大学 | Reading method for gate opening ruler |
| CN113819972A (en) * | 2020-07-07 | 2021-12-21 | 湖北亿立能科技股份有限公司 | Artificial intelligence surface of water detecting system based on virtual scale |
| CN115240394A (en) * | 2022-09-22 | 2022-10-25 | 国网湖北省电力有限公司经济技术研究院 | Method and system for monitoring and early warning water level of accident oil pool of transformer substation |
| CN116469091A (en) * | 2023-04-21 | 2023-07-21 | 浪潮智慧科技有限公司 | Automatic water gauge reading method, device and medium based on real-time video |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006047167A (en) * | 2004-08-06 | 2006-02-16 | Sankiki:Kk | Staff gage |
| CN101034154A (en) * | 2006-03-09 | 2007-09-12 | 华北电力大学 | Direct-reading instrument full-automatic verification method and device thereof |
| CN105741316A (en) * | 2016-01-20 | 2016-07-06 | 西北工业大学 | Robust target tracking method based on deep learning and multi-scale correlation filtering |
| CN105938554A (en) * | 2016-04-14 | 2016-09-14 | 北京交通大学 | Switch rail expansion displacement monitoring method based on image automatic interpretation and switch rail expansion displacement monitoring system thereof |
| CN105973342A (en) * | 2016-06-03 | 2016-09-28 | 湖南相水缘水利科技有限公司 | Intelligent water level monitoring system |
| CN106295566A (en) * | 2016-08-10 | 2017-01-04 | 北京小米移动软件有限公司 | Facial expression recognizing method and device |
| CN106599892A (en) * | 2016-12-14 | 2017-04-26 | 四川长虹电器股份有限公司 | Television station logo identification system based on deep learning |
| CN107038751A (en) * | 2015-12-07 | 2017-08-11 | 达索***公司 | The identification of 3D modeling object is carried out from 2D images |
| CN107122806A (en) * | 2017-05-16 | 2017-09-01 | 北京京东尚科信息技术有限公司 | A kind of nude picture detection method and device |
| CN107367310A (en) * | 2017-07-11 | 2017-11-21 | 华南理工大学 | A kind of river level remote monitoring method based on computer vision |
| CN107506798A (en) * | 2017-08-31 | 2017-12-22 | 福建四创软件有限公司 | A kind of water level monitoring method based on image recognition |
| CN107607207A (en) * | 2017-08-22 | 2018-01-19 | 中国科学院深圳先进技术研究院 | A kind of power equipment thermal fault detection method, system and electronic equipment |
-
2018
- 2018-04-10 CN CN201810315600.5A patent/CN108596221B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006047167A (en) * | 2004-08-06 | 2006-02-16 | Sankiki:Kk | Staff gage |
| CN101034154A (en) * | 2006-03-09 | 2007-09-12 | 华北电力大学 | Direct-reading instrument full-automatic verification method and device thereof |
| CN107038751A (en) * | 2015-12-07 | 2017-08-11 | 达索***公司 | The identification of 3D modeling object is carried out from 2D images |
| CN105741316A (en) * | 2016-01-20 | 2016-07-06 | 西北工业大学 | Robust target tracking method based on deep learning and multi-scale correlation filtering |
| CN105938554A (en) * | 2016-04-14 | 2016-09-14 | 北京交通大学 | Switch rail expansion displacement monitoring method based on image automatic interpretation and switch rail expansion displacement monitoring system thereof |
| CN105973342A (en) * | 2016-06-03 | 2016-09-28 | 湖南相水缘水利科技有限公司 | Intelligent water level monitoring system |
| CN106295566A (en) * | 2016-08-10 | 2017-01-04 | 北京小米移动软件有限公司 | Facial expression recognizing method and device |
| CN106599892A (en) * | 2016-12-14 | 2017-04-26 | 四川长虹电器股份有限公司 | Television station logo identification system based on deep learning |
| CN107122806A (en) * | 2017-05-16 | 2017-09-01 | 北京京东尚科信息技术有限公司 | A kind of nude picture detection method and device |
| CN107367310A (en) * | 2017-07-11 | 2017-11-21 | 华南理工大学 | A kind of river level remote monitoring method based on computer vision |
| CN107607207A (en) * | 2017-08-22 | 2018-01-19 | 中国科学院深圳先进技术研究院 | A kind of power equipment thermal fault detection method, system and electronic equipment |
| CN107506798A (en) * | 2017-08-31 | 2017-12-22 | 福建四创软件有限公司 | A kind of water level monitoring method based on image recognition |
Non-Patent Citations (3)
| Title |
|---|
| Y.-M.CHIANG等: "Dynamic neural networks for real-time water level predictions of sewerage systems – covering gauged and ungauged sites", 《HYDROLOGY AND EARTH SYSTEM SCIENCES DISCUSSIONS》 * |
| 张阳: "基于计算机视觉的河流水位测量***设计", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 陈金水: "基于视频图像识别的水位数据获取方法", 《水利信息化》 * |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109214806A (en) * | 2018-11-20 | 2019-01-15 | 北京京东尚科信息技术有限公司 | Self-help settlement method, apparatus and storage medium |
| US11861584B2 (en) | 2018-11-20 | 2024-01-02 | Beijing Jingdong Shangke Information Technology Co, Ltd. | Self-service settlement method, apparatus and storage medium |
| CN109784259A (en) * | 2019-01-08 | 2019-05-21 | 江河瑞通（北京）技术有限公司 | Water transparency intelligent identification Method and Sai Shi disk component based on image recognition |
| CN109784259B (en) * | 2019-01-08 | 2021-04-13 | 江河瑞通（北京）技术有限公司 | Intelligent water transparency identification method based on image identification and Samsung disk assembly |
| CN109859183B (en) * | 2019-01-29 | 2021-06-04 | 江河瑞通（北京）技术有限公司 | Edge calculation-based multi-element integrated water body intelligent identification method and ecological station |
| CN109859183A (en) * | 2019-01-29 | 2019-06-07 | 江河瑞通（北京）技术有限公司 | More element integrated water body intelligent identification Methods and ecology station based on edge calculations |
| CN110223341A (en) * | 2019-06-14 | 2019-09-10 | 北京国信华源科技有限公司 | A kind of Intelligent water level monitoring method based on image recognition |
| CN110223341B (en) * | 2019-06-14 | 2024-05-28 | 北京国信华源科技有限公司 | Intelligent water level monitoring method based on image recognition |
| CN111877105A (en) * | 2020-06-16 | 2020-11-03 | 中交第三公路工程局有限公司 | Modified asphalt construction method |
| CN113819972A (en) * | 2020-07-07 | 2021-12-21 | 湖北亿立能科技股份有限公司 | Artificial intelligence surface of water detecting system based on virtual scale |
| CN113112514A (en) * | 2021-04-27 | 2021-07-13 | 汇鸿智能科技(辽宁)有限公司 | Method and device for AI (Artificial Intelligence) recognition of graphite size, computer equipment and storage medium |
| CN113112514B (en) * | 2021-04-27 | 2024-05-17 | 汇鸿智能科技(辽宁)有限公司 | Method, device, computer equipment and storage medium for identifying graphite size by AI |
| CN113776408A (en) * | 2021-09-13 | 2021-12-10 | 北京邮电大学 | Reading method for gate opening ruler |
| CN115240394A (en) * | 2022-09-22 | 2022-10-25 | 国网湖北省电力有限公司经济技术研究院 | Method and system for monitoring and early warning water level of accident oil pool of transformer substation |
| CN116469091A (en) * | 2023-04-21 | 2023-07-21 | 浪潮智慧科技有限公司 | Automatic water gauge reading method, device and medium based on real-time video |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108596221B (en) | 2020-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108596221A (en) | The image-recognizing method and equipment of rod reading | |
| KR102423561B1 (en) | Intelligent dam management system based on digital twin | |
| JP6065427B2 (en) | Object tracking method and object tracking apparatus | |
| CN109815865B (en) | Water level identification method and system based on virtual water gauge | |
| CN108647574A (en) | Floating material image detection model generating method, recognition methods and equipment | |
| CN110458095A (en) | A kind of recognition methods, control method, device and the electronic equipment of effective gesture | |
| CN109344717B (en) | Multi-threshold dynamic statistical deep sea target online detection and identification method | |
| CN108009529B (en) | Forest fire smoke video target detection method based on characteristic root and hydrodynamics | |
| CN104933738B (en) | A kind of visual saliency map generation method detected based on partial structurtes with contrast | |
| CN108564069A (en) | A kind of industry safe wearing cap video detecting method | |
| den Bieman et al. | Deep learning video analysis as measurement technique in physical models | |
| CN101986348A (en) | Visual target identification and tracking method | |
| CN109886947A (en) | The high-tension bus-bar defect inspection method of convolutional neural networks based on region | |
| CN109580004A (en) | A kind of temperature checking method and device | |
| CN106960210A (en) | The method and apparatus of target detection | |
| CN116258608B (en) | Water conservancy real-time monitoring information management system integrating GIS and BIM three-dimensional technology | |
| CN103577875A (en) | CAD (computer-aided design) people counting method based on FAST (features from accelerated segment test) | |
| CN113177469A (en) | Training method and device for human body attribute detection model, electronic equipment and medium | |
| CN110276293A (en) | Method for detecting lane lines, device, electronic equipment and storage medium | |
| CN113642474A (en) | Hazardous area personnel monitoring method based on YOLOV5 | |
| CN110120065A (en) | A kind of method for tracking target and system based on layering convolution feature and dimension self-adaption core correlation filtering | |
| CN107622503A (en) | A kind of layering dividing method for recovering image Ouluding boundary | |
| CN109840905A (en) | Power equipment rusty stain detection method and system | |
| CN109583393A (en) | A kind of lane line endpoints recognition methods and device, equipment, medium | |
| CN115797846A (en) | Wind power generation blade block defect comparison method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address |
Address after: 100097 room 10e, unit 2 (block B), 9 / F, building 2, yard 2, lantianchang East Road, Haidian District, Beijing (Building 2, Jinyuan times business center) Patentee after: Kunyu (Beijing) Technology Co.,Ltd. Address before: A646, 7th Floor, 7th Floor, No. 28 Xinxin Road, Haidian District, Beijing, 100085 Patentee before: RICHWAY (BEIJING) TECHNOLOGY Co.,Ltd. |
|
| CP03 | Change of name, title or address |