CN111815623A - Power transmission line cotter pin missing identification method - Google Patents
Power transmission line cotter pin missing identification method Download PDFInfo
- Publication number
- CN111815623A CN111815623A CN202010737726.9A CN202010737726A CN111815623A CN 111815623 A CN111815623 A CN 111815623A CN 202010737726 A CN202010737726 A CN 202010737726A CN 111815623 A CN111815623 A CN 111815623A
- Authority
- CN
- China
- Prior art keywords
- component
- target
- missing
- frame
- detected
- 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
- 230000005540 biological transmission Effects 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000001514 detection method Methods 0.000 claims abstract description 167
- 238000007689 inspection Methods 0.000 claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 10
- 230000015654 memory Effects 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 6
- 230000005764 inhibitory process Effects 0.000 claims description 2
- 230000007547 defect Effects 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 230000001629 suppression Effects 0.000 description 4
- 239000000725 suspension Substances 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000002950 deficient Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- 239000002184 metal Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000009966 trimming Methods 0.000 description 2
- 238000004873 anchoring Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4084—Scaling of whole images or parts thereof, e.g. expanding or contracting in the transform domain, e.g. fast Fourier transform [FFT] domain scaling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Quality & Reliability (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The application relates to a method for identifying missing of a cotter pin of a power transmission line. The method for identifying the missing of the cotter pin of the power transmission line comprises the steps of obtaining a patrol picture of the power transmission line and generating a sample set. And training the sample set to respectively obtain a component detection model and a cotter pin missing detection model. And inputting the inspection pictures to be detected into the component detection model to generate a plurality of frames to be detected of the components. And carrying out hierarchical division on target parts in the plurality of parts to-be-detected frames to determine a part detection frame. And detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part. According to the method, the target component is subjected to hierarchical division to remove false detection and missed detection, and the detection accuracy is improved.
Description
Technical Field
The application relates to the technical field of detection, in particular to a method for identifying missing of a cotter pin of a power transmission line.
Background
Along with the large-scale popularization of unmanned aerial vehicle inspection business of the power transmission line, a large number of inspection pictures can be generated every day, and the pictures with the defects also have the defects, so that the defects need to be picked out by using an intelligent identification technology, the types of the defects are identified, and the positions of the defects are located. The split pin mainly prevents that the bolt from droing to lead to junction or hang some disconnection, so that transmission line and tower overall structure's stability suffers destruction. Difficulty of intelligent identification of cotter: 1. the target is small, the cotter pin is small relative to the original drawing, and the deficiency is searched just like a large sea fishing needle; 2. the interference is much, and a large number of bolts which do not need to be provided with cotter pins are distributed on the parts such as the tower, hardware fittings and the like; 3. the distribution is wide, and the bolts for installing the split pins are distributed at hanging points, connection, anchoring and other places; 4. the illumination is uneven, the light below the tower is too dark, and the light is too strong in sunny days; 5. fuzzy, the operation of the machine patrol personnel has the problems of motion fuzzy and focus fuzzy.
At present, the academic world and the industrial world do not have good solutions for intelligent identification of small targets with unfixed environments and many interference factors, such as cotter pin missing, mainly analyze the small targets by means of manual checking, and since the original pictures are large and the defective targets are small, the pictures need to be frequently enlarged and reduced to find whether defects exist, so that the workload is large, the efficiency is low, and the defective pictures are easy to miss.
Disclosure of Invention
Based on this, the application provides a transmission line cotter pin missing identification method aiming at the problems of more missed detections, high false detections and the like in the traditional technical scheme.
A method for identifying missing of a cotter pin of a power transmission line comprises the following steps:
acquiring a patrol picture of the power transmission line, and generating a sample set;
training the sample set to respectively obtain a component detection model and a cotter missing detection model;
inputting a to-be-detected inspection picture into the component detection model to generate a plurality of component detection frames, wherein the component detection frames comprise target components;
performing hierarchical division on target parts in the plurality of parts to-be-detected frames to determine part detection frames;
and detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part.
In one embodiment, the step of inputting the inspection picture to be detected into the component inspection model further comprises the following steps:
and cutting the to-be-detected inspection picture to avoid missing inspection of the target component.
In one embodiment, the step of inputting the inspection picture to be detected into the component detection model to generate a plurality of frames to be detected of the component includes:
and inputting the to-be-inspected picture and the cut picture into the component detection model to obtain the position information of the target component in each component to-be-inspected frame and the confidence coefficient of the target component.
In one embodiment, the step of inputting the inspection picture to be detected into the component detection model to generate a plurality of frames to be detected of the component, where the step of including the target component in the component detection frame includes the following steps:
removing the part to-be-detected frame of which the confidence coefficient of the target part is smaller than a first threshold value;
and removing the part inspection frame with the overlapping area larger than the second threshold value by using a non-maximum value inhibition algorithm.
In one embodiment, the step of inputting the inspection picture to be detected into the component detection model to generate a plurality of frames to be detected for the component, wherein the step of including the target component in the component detection frame further includes:
judging whether the size of the target part in the part frame to be detected is smaller than a fifth threshold value according to the position information of the target part in the part frame to be detected;
and when the size of the target part in the part inspection frame is smaller than a fifth threshold value, removing the current part inspection frame.
In one embodiment, the step of hierarchically dividing the target component in the component checkbox to determine the component detection box includes:
dividing the target component into a primary target component and a secondary target component, wherein the size of the primary target component is larger than the size of the secondary target component;
when the part frame to be detected comprises a first-stage target part and a second-stage target part, judging whether the second-stage target part is positioned in the first-stage target part;
and when the secondary target component is positioned in the primary target component, removing the frame to be detected of the current component detection frame component, and further determining the component detection frame.
In one embodiment, the step of detecting each component detection frame by using the missing cotter detection model to obtain the position information of the target component, the category information of the target component, and the confidence level of the target component includes:
normalizing the size of the component detection frame to ensure the scaling of the height and the width in equal proportion;
and detecting the normalized component detection frame to acquire the position information of the target component, the category information of the target component and the confidence coefficient of the target component.
In one embodiment, the step of detecting each component detection frame by using the missing cotter detection model to obtain the position information of the target component, the category information of the target component, and the confidence level of the target component is followed by:
and removing the cotter pin missing part detection frame with the confidence coefficient of the target part being smaller than a third threshold value.
In one embodiment, the step of detecting each component detection frame by using the missing cotter detection model to obtain the position information of the target component, the category information of the target component, and the confidence level of the target component is followed by:
determining whether a cotter-missing part detection frame falls on the secondary target part;
outputting a missing cotter pin when the cotter pin missing part detection frame falls on the secondary target part;
when the cotter missing part detection frame does not fall on the secondary target part, determining whether the confidence of the target part is greater than a fourth threshold;
outputting a cotter pin missing when the confidence of the target component is greater than the fourth threshold.
A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method of identifying a loss of a cotter pin in any one of the above embodiments when executing the computer program.
The method for identifying the missing of the cotter pin of the power transmission line comprises the steps of obtaining a patrol picture of the power transmission line and generating a sample set. And training the sample set to respectively obtain a component detection model and a cotter pin missing detection model. And inputting the inspection pictures to be detected into the component detection model to generate a plurality of frames to be detected of the components. And carrying out hierarchical division on target parts in the plurality of parts to-be-detected frames to determine a part detection frame. And detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part. According to the method, the target component is subjected to hierarchical division to remove false detection and missed detection, and the detection accuracy is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments or the conventional technologies of the present application, the drawings used in the descriptions of the embodiments or the conventional technologies will be briefly introduced below, it is obvious that the drawings in the following descriptions are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flowchart of a method for identifying missing cotter pins of a power transmission line according to an embodiment of the present application;
fig. 2 is a schematic flow chart of a method for identifying missing cotter pins of a power transmission line according to another embodiment of the present application.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present application more comprehensible, embodiments accompanying the present application are described in detail below with reference to the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is capable of embodiments in many different forms than those described herein and those skilled in the art will be able to make similar modifications without departing from the spirit of the application and it is therefore not intended to be limited to the embodiments disclosed below.
It will be understood that, as used herein, the terms "first," "second," and the like may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one element from another. For example, a first acquisition module may be referred to as a second acquisition module, and similarly, a second acquisition module may be referred to as a first acquisition module, without departing from the scope of the present application. The first acquisition module and the second acquisition module are both acquisition modules, but are not the same acquisition module.
It will be understood that when an element is referred to as being "disposed on" another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
Referring to fig. 1, the present application provides a method for identifying a missing cotter pin of a power transmission line. The method for identifying the missing of the cotter pin of the power transmission line comprises the following steps:
and S10, acquiring the inspection picture of the power transmission line and generating a sample set.
In step S10, the manner of acquiring the sample is not particularly limited. In an alternative embodiment, images of the unmanned aerial vehicle in various states of the power transmission line and the tower can be used as samples. The images in all the states mainly comprise images in different angles, different states, different shapes and different environments, so that the trained model has stronger adaptability.
And S20, training the sample set to respectively obtain a component detection model and a cotter missing detection model.
In step S20, the sample may be trained based on the Cascade RCNN + FPN + ResNet50 algorithm to obtain a component detection model. The samples can be trained based on the Cascade RCNN + Resnet50+ FPN + DCNv2 algorithm to obtain a cotter pin loss detection model.
The Cascade RCNN algorithm achieves the purpose of continuously optimizing the prediction result by cascading a plurality of detection networks, and is different from the common Cascade, wherein the plurality of detection networks of the Cascade RCNN are obtained by training positive and negative samples determined based on different IOU thresholds.
And S30, inputting the inspection pictures to be detected into the component detection model, and generating a plurality of component frames to be detected, wherein the component detection frames contain target components.
In step S30, the target component may be a hanging point fitting, a link fitting, a suspension clamp, a tie clamp, a drainage plate, a U-shaped hanging ring, a chain, a long clamping piece, a short clamping piece, or a right-angle hanging plate. Referring to fig. 2, since the parts and components occupy too little in the 8K image, directly detecting and identifying the parts on the original image may cause many object missing detections, so that step S30 may further include cropping the inspection picture to be detected to avoid missing the object parts. The artwork may be cropped using a fixed size 1640 x 1000 with the sliding window overlapping 200 pixels.
In order to prevent missing inspection or incomplete inspection of the component detection frame caused by incomplete trimming of the target component with a larger size during trimming, the step S30 may further include inputting both the inspection picture to be inspected and the trimmed picture into the component detection model, and obtaining position information of the target component in the inspection frame of each component and confidence of the target component.
S40, performing hierarchical division on the target parts in the parts to-be-detected frames to determine the parts detection frames.
In step S40, in order to reduce false detection and reduce the calculation pressure, the generated large number of part checkboxes may be subjected to false detection removal and part checkbox fusion before being subjected to hierarchical division. Optionally, the step of false-detecting may be removing the part-to-be-detected box of which the confidence of the target part is smaller than the first threshold. The first threshold may be 0.5. Optionally, the step of merging the part checkboxes may be to remove the part checkboxes with the overlapping area larger than the second threshold using a non-maximum suppression algorithm. In the metal detection process, a plurality of parts to be detected frames and confidence scores thereof are obtained. And removing the redundant part inspection frame with large coincidence rate by using a non-maximum suppression algorithm to obtain a most representative result so as to accelerate the target detection efficiency. The second threshold may be 0.2.
In one optional embodiment, before the step of performing hierarchical division, it may be further determined whether the size of the target part in the part checkbox is smaller than a fifth threshold according to the position information of the target part in the part checkbox. And when the size of the target part in the part inspection frame is smaller than a fifth threshold value, removing the current part inspection frame. The size of the fifth threshold is not particularly limited. For the part to-be-detected frame comprising the part and the part with smaller sizes, the part and the part with smaller sizes are fuzzy and difficult to distinguish by human eyes, and whether the cotter pin on the part is normal or not can not be distinguished, so that the part to-be-detected frame can be filtered by adopting relative sizes and absolute sizes. The absolute magnitude filtering removes the ambiguity by a fixed length-to-width threshold (220 x 220). Relative size filtering by alignment with the largest target of the same type, areas less than 1/4 were filtered.
In one of the alternative embodiments, in order to reduce the amount of calculation, a determination may be made as to whether the hardware component performs pin missing detection, so as to ensure that only the "large" component and the "medium" part that does not fall inside the large component perform identification of the missing cotter pin. Therefore, the size information of the target component can be obtained according to the position information of the target component, and the target component is further divided into a first-stage target component and a second-stage target component. And when the component detection frame comprises a primary target component and a secondary target component, performing no cotter pin missing detection on the secondary target component positioned in the primary target component in the component detection frame. The primary target component is a large component such as a connecting hardware fitting, a hanging point hardware fitting, a suspension clamp, a U-shaped ring, a drainage plate and a pull clamp. The secondary target parts are middle-sized and middle-sized parts such as chains, short clamping pieces, long clamping pieces, right-angle hanging plates and the like.
And S50, detecting each component detection frame by using the cotter missing detection model to acquire the position information of the target component, the category information of the target component and the confidence coefficient of the target component.
In step S50, the position and the confidence level information of the normal cotter pin or the position and the confidence level information of the missing cotter pin can be determined from the position information of the target component, the category information of the target component, and the confidence level of the target component. In one embodiment, step S50 includes:
and normalizing the size of the component detection frame to ensure the scaling of the height and the width. And detecting the normalized component detection frame to acquire the position information of the target component, the category information of the target component and the confidence coefficient of the target component. In particular, the component detection box may be scaled uniformly to a dimension of 1312 × 800, while ensuring that the height and width are scaled uniformly, filling the edges with [104, 117, 123] pixel values. And detecting the 'small' original (the cotter pin is normal or the cotter pin is missing) of the normalized 'large' part and 'medium' part detection frames by using the cotter pin missing detection model.
In this embodiment, the method for identifying the missing of the cotter pin of the power transmission line includes obtaining a patrol picture of the power transmission line and generating a sample set. And training the sample set to respectively obtain a component detection model and a cotter pin missing detection model. And inputting the inspection pictures to be detected into the component detection model to generate a plurality of frames to be detected of the components. And carrying out hierarchical division on target parts in the plurality of parts to-be-detected frames to determine a part detection frame. And detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part. According to the method, the target component is subjected to hierarchical division to remove false detection and missed detection, and the detection accuracy is improved.
In one embodiment, step S50 is followed by removing the cotter missing part detection box with the confidence of the target part being less than a third threshold. The third threshold may be 0.8. The step can further improve the detection precision and reduce the false detection rate.
In one embodiment, step S50 is followed by:
it is determined whether the cotter-missing part detection frame falls on the secondary target part. Outputting a cotter missing when the cotter missing part detection frame falls on the secondary target part. When the cotter-missing part detection frame does not fall on the secondary target part, it is determined whether the confidence of the target part is greater than a fourth threshold. Outputting a cotter pin missing when the confidence of the target component is greater than the fourth threshold.
In this embodiment, whether the cotter pin missing part detection frame is actually missing is determined by determining whether the cotter pin missing part detection frame falls on the "middle" part, that is, if the cotter pin missing part detection frame falls on the "middle" part, the cotter pin missing part detection frame is determined to be actually missing, otherwise, the cotter pin missing part detection frame is false detection. According to the method and the device, the false detection is removed through the logic judgment of the hierarchical relationship, and in addition, whether the missing of the detected cotter pin falls on the middle part or not is judged, so that the missing identification of the pin is realized, and the detection rate of the missing is improved.
In one embodiment of the present application, a computer device is provided. The computer device comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the computer program to realize the steps of the method for identifying the missing of the cotter pin of the power transmission line in any one of the above embodiments.
The memory is a computer-readable storage medium, and may be used to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the method for identifying a missing cotter pin in an electric transmission line in the embodiment of the present application. The processor executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory, namely, the method for identifying the loss of the cotter pin of the power transmission line is realized.
The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function. The storage data area may store data created according to the use of the terminal, and the like. Further, the memory may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, the memory may further include memory located remotely from the processor, and these remote memories may be connected to the device over a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The processor can acquire the inspection picture of the power transmission line by running the software program, the instruction and the module stored in the memory to generate a sample set; training the sample set to respectively obtain a component detection model and a cotter missing detection model; inputting a to-be-detected inspection picture into the component detection model to generate a plurality of component detection frames, wherein the component detection frames comprise target components; performing hierarchical division on target parts in the plurality of parts to-be-detected frames to determine part detection frames; and detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part.
The manner of obtaining the sample is not particularly limited. In an alternative embodiment, images of the unmanned aerial vehicle in various states of the power transmission line and the tower can be used as samples. The images in all the states mainly comprise images in different angles, different states, different shapes and different environments, so that the trained model has stronger adaptability.
The samples may be trained based on the Cascade RCNN + FPN + ResNet50 algorithm to obtain a component inspection model. The samples can be trained based on the Cascade RCNN + Resnet50+ FPN + DCNv2 algorithm to obtain a cotter pin loss detection model.
The Cascade RCNN algorithm achieves the purpose of continuously optimizing the prediction result by cascading a plurality of detection networks, and is different from the common Cascade, wherein the plurality of detection networks of the Cascade RCNN are obtained by training positive and negative samples determined based on different IOU thresholds.
The target component can be a hanging point hardware fitting, a connecting hardware fitting, a suspension wire clamp, a pull wire clamp, a drainage plate, a U-shaped hanging ring, a chain, a long clamping piece, a short clamping piece and a right-angle hanging plate. Referring to fig. 2, since the parts and components occupy too little in the 8K image, and detection and identification of the parts and components directly on the original image may cause many objects to be missed, the processor may also implement cropping of the inspection picture to be detected by running the software program, the instructions, and the modules stored in the memory, so as to avoid the step of missing the objects. The artwork may be cropped using a fixed size 1640 x 1000 with the sliding window overlapping 200 pixels.
In order to prevent missing detection or incomplete detection of the component caused by incomplete cutting of a target component with a large size during cutting, the processor can also realize the step of inputting the inspection picture to be detected and the cut picture into the component detection model by running a software program, an instruction and a module which are stored in a memory, and obtaining the position information of the target component and the confidence coefficient of the target component in each component frame to be detected.
In order to reduce false detection and reduce calculation pressure, a large number of generated part checkboxes can be subjected to false detection removal and part checkbox fusion before hierarchical division. Optionally, the step of false-detecting may be removing the part-to-be-detected box of which the confidence of the target part is smaller than the first threshold. The first threshold may be 0.5. Optionally, the step of merging the part checkboxes may be to remove the part checkboxes with the overlapping area larger than the second threshold using a non-maximum suppression algorithm. In the metal detection process, a plurality of parts to be detected frames and confidence scores thereof are obtained. And removing the redundant part inspection frame with large coincidence rate by using a non-maximum suppression algorithm to obtain a most representative result so as to accelerate the target detection efficiency. The second threshold may be 0.2.
In one optional embodiment, before the step of performing hierarchical division, it may be further determined whether the size of the target part in the part checkbox is smaller than a fifth threshold according to the position information of the target part in the part checkbox. And when the size of the target part in the part inspection frame is smaller than a fifth threshold value, removing the current part inspection frame. The size of the fifth threshold is not particularly limited. For the part to-be-detected frame comprising the part and the part with smaller sizes, the part and the part with smaller sizes are fuzzy and difficult to distinguish by human eyes, and whether the cotter pin on the part is normal or not can not be distinguished, so that the part to-be-detected frame can be filtered by adopting relative sizes and absolute sizes. The absolute magnitude filtering removes the ambiguity by a fixed length-to-width threshold (220 x 220). Relative size filtering by alignment with the largest target of the same type, areas less than 1/4 were filtered.
In one of the alternative embodiments, in order to reduce the amount of calculation, a determination may be made as to whether the hardware component performs pin missing detection, so as to ensure that only the "large" component and the "medium" part that does not fall inside the large component perform identification of the missing cotter pin. Therefore, the size information of the target component can be obtained according to the position information of the target component, and the target component is further divided into a first-stage target component and a second-stage target component. And when the component detection frame comprises a primary target component and a secondary target component, performing no cotter pin missing detection on the secondary target component positioned in the primary target component in the component detection frame. The primary target component is a large component such as a connecting hardware fitting, a hanging point hardware fitting, a suspension clamp, a U-shaped ring, a drainage plate and a pull clamp. The secondary target parts are middle-sized and middle-sized parts such as chains, short clamping pieces, long clamping pieces, right-angle hanging plates and the like.
And determining the position and the confidence degree information of the normal cotter pin or the missing position and the confidence degree information of the cotter pin according to the position information of the target component, the category information of the target component and the confidence degree of the target component.
In one embodiment, the processor, by executing the software program, instructions and modules stored in the memory, may further perform normalization of the dimensions of the component detection box to ensure aspect scaling. And detecting the normalized component detection frame to acquire the position information of the target component, the category information of the target component and the confidence coefficient of the target component. In particular, the component detection box may be scaled uniformly to a dimension of 1312 × 800, while ensuring that the height and width are scaled uniformly, filling the edges with [104, 117, 123] pixel values. And detecting the 'small' original (the cotter pin is normal or the cotter pin is missing) of the normalized 'large' part and 'medium' part detection frames by using the cotter pin missing detection model.
The computer equipment comprises a polling picture for acquiring the power transmission line and generates a sample set. And training the sample set to respectively obtain a component detection model and a cotter pin missing detection model. And inputting the inspection pictures to be detected into the component detection model to generate a plurality of frames to be detected of the components. And carrying out hierarchical division on target parts in the plurality of parts to-be-detected frames to determine a part detection frame. And detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part. According to the method, the target component is subjected to hierarchical division to remove false detection and missed detection, and the detection accuracy is improved.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the claims. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method for identifying the missing of a cotter pin of a power transmission line is characterized by comprising the following steps:
acquiring a patrol picture of the power transmission line, and generating a sample set;
training the sample set to respectively obtain a component detection model and a cotter missing detection model;
inputting a to-be-detected inspection picture into the component detection model to generate a plurality of component detection frames, wherein the component detection frames comprise target components;
performing hierarchical division on a target part in the part to-be-detected frame to determine a part detection frame;
and detecting each part detection frame by using the cotter missing detection model to acquire the position information of the target part, the category information of the target part and the confidence coefficient of the target part.
2. The method for identifying the missing cotter pin of the power transmission line according to claim 1, wherein the step of inputting the inspection picture to be detected into the component detection model further comprises the following steps:
and cutting the to-be-detected inspection picture to avoid missing inspection of the target component.
3. The method for identifying the missing cotter pin of the power transmission line according to claim 2, wherein the step of inputting the inspection pictures to be detected into the component detection model to generate the frames to be detected of the plurality of components comprises the following steps:
and inputting the to-be-inspected picture and the cut picture into the component detection model to obtain the position information of the target component in each component to-be-inspected frame and the confidence coefficient of the target component.
4. The method for identifying the missing cotter pin of the power transmission line according to claim 3, wherein the step of inputting the inspection pictures to be detected into the component detection model to generate a plurality of frames to be detected of the component, and the step of including the target component in the component detection frame comprises the following steps:
removing the part to-be-detected frame of which the confidence coefficient of the target part is smaller than a first threshold value;
and removing the part inspection frame with the overlapping area larger than the second threshold value by using a non-maximum value inhibition algorithm.
5. The method for identifying the missing cotter pin of the power transmission line according to claim 4, wherein the inspection picture to be detected is input into the component detection model to generate a plurality of frames to be detected for the component, and the steps of including the target component in the component detection frame further comprise:
judging whether the size of the target part in the part frame to be detected is smaller than a fifth threshold value according to the position information of the target part in the part frame to be detected;
and when the size of the target part in the part inspection frame is smaller than a fifth threshold value, removing the current part inspection frame.
6. The method for identifying the missing cotter pin of the power transmission line according to claim 5, wherein the step of performing hierarchical division on the target part in the frame to be inspected to determine the part detection frame comprises:
dividing the target component into a primary target component and a secondary target component, wherein the size of the primary target component is larger than the size of the secondary target component;
when the part frame to be detected comprises a first-stage target part and a second-stage target part, judging whether the second-stage target part is positioned in the first-stage target part;
and when the secondary target component is positioned in the primary target component, removing the frame to be detected of the current component detection frame component, and further determining the component detection frame.
7. The method according to claim 6, wherein the step of detecting each component detection frame by using the cotter missing detection model to obtain the position information of the target component, the category information of the target component and the confidence level of the target component comprises:
normalizing the size of the component detection frame to ensure the scaling of the height and the width in equal proportion;
and detecting the normalized component detection frame to acquire the position information of the target component, the category information of the target component and the confidence coefficient of the target component.
8. The method according to claim 7, wherein the step of detecting each component detection frame by using the cotter missing detection model to obtain the position information of the target component, the category information of the target component, and the confidence level of the target component comprises the following steps:
and removing the cotter pin missing part detection frame with the confidence coefficient of the target part being smaller than a third threshold value.
9. The method according to claim 8, wherein the step of detecting each component detection frame by using the missing cotter detection model to obtain the position information of the target component, the category information of the target component, and the confidence level of the target component comprises the steps of:
determining whether a cotter-missing part detection frame falls on the secondary target part;
outputting a missing cotter pin when the cotter pin missing part detection frame falls on the secondary target part;
when the cotter missing part detection frame does not fall on the secondary target part, determining whether the confidence of the target part is greater than a fourth threshold;
outputting a cotter pin missing when the confidence of the target component is greater than the fourth threshold.
10. Computer arrangement comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor when executing the computer program realizes the steps of the method for identifying the absence of a cotter pin of a power transmission line according to any one of claims 1 to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010737726.9A CN111815623B (en) | 2020-07-28 | 2020-07-28 | Power transmission line cotter pin missing identification method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010737726.9A CN111815623B (en) | 2020-07-28 | 2020-07-28 | Power transmission line cotter pin missing identification method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111815623A true CN111815623A (en) | 2020-10-23 |
CN111815623B CN111815623B (en) | 2024-02-23 |
Family
ID=72863215
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010737726.9A Active CN111815623B (en) | 2020-07-28 | 2020-07-28 | Power transmission line cotter pin missing identification method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111815623B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112312025A (en) * | 2020-10-30 | 2021-02-02 | 南方电网数字电网研究院有限公司 | Target deviation rectifying method and system based on double zoom holder |
CN112907561A (en) * | 2021-03-16 | 2021-06-04 | 中科海拓(无锡)科技有限公司 | Notebook appearance flaw detection method based on deep learning |
CN112926401A (en) * | 2021-01-29 | 2021-06-08 | 广州中科智巡科技有限公司 | Transmission line hardware corrosion detection method and system |
CN114120164A (en) * | 2021-09-28 | 2022-03-01 | 佛山中科云图智能科技有限公司 | Power transmission line pin state detection method and detection system |
CN115731478A (en) * | 2022-11-24 | 2023-03-03 | 国网湖北省电力有限公司超高压公司 | Power transmission line cotter pin target detection method based on multistage target detection |
CN116012375A (en) * | 2023-03-22 | 2023-04-25 | 成都唐源电气股份有限公司 | Method and system for detecting cotter pin defects of overhead contact system soft crossing suspension pulley |
CN113744211B (en) * | 2021-08-19 | 2023-11-03 | 衢州光明电力投资集团有限公司赋腾科技分公司 | Target part missing detection method based on azimuth combination in image |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106226050A (en) * | 2016-07-15 | 2016-12-14 | 北京航空航天大学 | A kind of TFDS fault automatic identifying method |
CN106951899A (en) * | 2017-02-24 | 2017-07-14 | 李刚毅 | Method for detecting abnormality based on image recognition |
CN108389197A (en) * | 2018-02-26 | 2018-08-10 | 上海赛特斯信息科技股份有限公司 | Transmission line of electricity defect inspection method based on deep learning |
CN109977817A (en) * | 2019-03-14 | 2019-07-05 | 南京邮电大学 | EMU car bed bolt fault detection method based on deep learning |
CN110232687A (en) * | 2019-06-21 | 2019-09-13 | 华北电力大学(保定) | The detection method of cotter bolt defect in a kind of electric inspection process image |
CN110246132A (en) * | 2019-06-23 | 2019-09-17 | 中车青岛四方车辆研究所有限公司 | Rail vehicle bolt looseness detection method and system |
CN110569762A (en) * | 2019-08-27 | 2019-12-13 | 许昌许继软件技术有限公司 | pin falling detection method and device based on multistage neural network |
CN110942455A (en) * | 2019-11-26 | 2020-03-31 | 南方电网数字电网研究院有限公司 | Method and device for detecting missing of cotter pin of power transmission line and computer equipment |
CN111127416A (en) * | 2019-12-19 | 2020-05-08 | 武汉珈鹰智能科技有限公司 | Computer vision-based automatic detection method for surface defects of concrete structure |
US20200175352A1 (en) * | 2017-03-14 | 2020-06-04 | University Of Manitoba | Structure defect detection using machine learning algorithms |
CN111372042A (en) * | 2020-01-16 | 2020-07-03 | 上海眼控科技股份有限公司 | Fault detection method and device, computer equipment and storage medium |
CN111444809A (en) * | 2020-03-23 | 2020-07-24 | 华南理工大学 | Power transmission line abnormal target detection method based on improved YO L Ov3 |
-
2020
- 2020-07-28 CN CN202010737726.9A patent/CN111815623B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106226050A (en) * | 2016-07-15 | 2016-12-14 | 北京航空航天大学 | A kind of TFDS fault automatic identifying method |
CN106951899A (en) * | 2017-02-24 | 2017-07-14 | 李刚毅 | Method for detecting abnormality based on image recognition |
US20200175352A1 (en) * | 2017-03-14 | 2020-06-04 | University Of Manitoba | Structure defect detection using machine learning algorithms |
CN108389197A (en) * | 2018-02-26 | 2018-08-10 | 上海赛特斯信息科技股份有限公司 | Transmission line of electricity defect inspection method based on deep learning |
CN109977817A (en) * | 2019-03-14 | 2019-07-05 | 南京邮电大学 | EMU car bed bolt fault detection method based on deep learning |
CN110232687A (en) * | 2019-06-21 | 2019-09-13 | 华北电力大学(保定) | The detection method of cotter bolt defect in a kind of electric inspection process image |
CN110246132A (en) * | 2019-06-23 | 2019-09-17 | 中车青岛四方车辆研究所有限公司 | Rail vehicle bolt looseness detection method and system |
CN110569762A (en) * | 2019-08-27 | 2019-12-13 | 许昌许继软件技术有限公司 | pin falling detection method and device based on multistage neural network |
CN110942455A (en) * | 2019-11-26 | 2020-03-31 | 南方电网数字电网研究院有限公司 | Method and device for detecting missing of cotter pin of power transmission line and computer equipment |
CN111127416A (en) * | 2019-12-19 | 2020-05-08 | 武汉珈鹰智能科技有限公司 | Computer vision-based automatic detection method for surface defects of concrete structure |
CN111372042A (en) * | 2020-01-16 | 2020-07-03 | 上海眼控科技股份有限公司 | Fault detection method and device, computer equipment and storage medium |
CN111444809A (en) * | 2020-03-23 | 2020-07-24 | 华南理工大学 | Power transmission line abnormal target detection method based on improved YO L Ov3 |
Non-Patent Citations (4)
Title |
---|
ZHENBING ZHAO等: "Detection Method Based on Automatic Visual Shape Clustering for Pin-Missing Defect in Transmission Lines", 《IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT》, vol. 69, no. 9, pages 1 - 12 * |
侯鑫鑫等: "基于图像检测与分类级联模型的车辆刮痕识别", 《工业控制计算机》, vol. 32, no. 3, pages 15 - 16 * |
王子昊: "深度学习在输电铁塔关键部件缺陷检测中的应用研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》, no. 10, pages 26 - 30 * |
闵锋等: "基于语义分割的接触网开口销状态检测", 《华中科技大学学报(自然科学版)》, vol. 48, no. 1, pages 77 - 81 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112312025A (en) * | 2020-10-30 | 2021-02-02 | 南方电网数字电网研究院有限公司 | Target deviation rectifying method and system based on double zoom holder |
CN112926401A (en) * | 2021-01-29 | 2021-06-08 | 广州中科智巡科技有限公司 | Transmission line hardware corrosion detection method and system |
CN112907561A (en) * | 2021-03-16 | 2021-06-04 | 中科海拓(无锡)科技有限公司 | Notebook appearance flaw detection method based on deep learning |
CN113744211B (en) * | 2021-08-19 | 2023-11-03 | 衢州光明电力投资集团有限公司赋腾科技分公司 | Target part missing detection method based on azimuth combination in image |
CN114120164A (en) * | 2021-09-28 | 2022-03-01 | 佛山中科云图智能科技有限公司 | Power transmission line pin state detection method and detection system |
CN115731478A (en) * | 2022-11-24 | 2023-03-03 | 国网湖北省电力有限公司超高压公司 | Power transmission line cotter pin target detection method based on multistage target detection |
CN115731478B (en) * | 2022-11-24 | 2023-12-22 | 国网湖北省电力有限公司超高压公司 | Power transmission line cotter pin target detection method based on multistage target detection |
CN116012375A (en) * | 2023-03-22 | 2023-04-25 | 成都唐源电气股份有限公司 | Method and system for detecting cotter pin defects of overhead contact system soft crossing suspension pulley |
Also Published As
Publication number | Publication date |
---|---|
CN111815623B (en) | 2024-02-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111815623A (en) | Power transmission line cotter pin missing identification method | |
US11488294B2 (en) | Method for detecting display screen quality, apparatus, electronic device and storage medium | |
US20200380899A1 (en) | Method and apparatus for detecting peripheral circuit of display screen, electronic device, and storage medium | |
CN110650316A (en) | Intelligent patrol and early warning processing method and device, electronic equipment and storage medium | |
US20200355627A1 (en) | Method for detecting display screen peripheral circuit, apparatus, electronic device and storage medium | |
CN110321933B (en) | Fault identification method and device based on deep learning | |
CN112560816A (en) | Equipment indicator lamp identification method and system based on YOLOv4 | |
CN105260716A (en) | Fault indicator state identification method and fault indicator state identification device | |
CN111695493A (en) | Method and system for detecting hidden danger of power transmission line | |
CN112330597A (en) | Image difference detection method and device and computer equipment | |
CN109724993A (en) | Detection method, device and the storage medium of the degree of image recognition apparatus | |
CN113326783A (en) | Edge early warning method for water conservancy industry | |
CN116503817A (en) | Street lamp illumination monitoring method, system, terminal, medium and illumination system based on image recognition | |
CN111464790A (en) | Industrial environment monitoring video processing method and system based on deep learning | |
CN111398292A (en) | Gabor filtering and CNN-based cloth defect detection method, system and equipment | |
CN111462074B (en) | Bearing appearance detection method, device, system, computer equipment and storage medium | |
CN113095225A (en) | System, method and equipment for artificial intelligent detection of operating state of pumping unit through infrared panoramic observation and storage medium | |
CN112967224A (en) | Electronic circuit board detection system, method and medium based on artificial intelligence | |
CN112001336A (en) | Pedestrian boundary crossing alarm method, device, equipment and system | |
CN111985497B (en) | Crane operation identification method and system under overhead transmission line | |
CN109860742B (en) | Method for identifying electrolyte leakage of communication power supply storage battery of transformer substation | |
CN108985222B (en) | Deep learning network model and system for recognition of incoming calls | |
CN113420631A (en) | Safety alarm method and device based on image recognition | |
CN112464928B (en) | Digital meter reading identification method, device, equipment and storage medium | |
CN115273013B (en) | Lane line detection method, system, computer and readable storage medium |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: Room 86, room 406, No.1, Yichuang street, Zhongxin Guangzhou Knowledge City, Huangpu District, Guangzhou City, Guangdong Province Applicant after: Southern Power Grid Digital Grid Research Institute Co.,Ltd. Address before: 511458 Room 1301, Chengtou Building, 106 Fengze East Road, Nansha District, Guangzhou City, Guangdong Province (self-compiled 1301-12159) Applicant before: Southern Power Grid Digital Grid Research Institute Co.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |