Abrasion detection method and system for rail transit contact line
Through image processing and feature point matching technology, the low efficiency and accuracy of contact line wear detection are solved, efficient and accurate wear evaluation and early warning are achieved, and manual workload and environmental factors are reduced.
Patent Information
- Application Number
- CN202510743160.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-05
- Publication Date
- 2025-09-02
AI Technical Summary
In the prior art, contact line wear detection relies on manual inspection and mechanical measuring instruments, and there are problems of low efficiency, easy leakage inspection and easy to be affected by environmental factors, making it difficult to achieve efficient and accurate wear assessment.
Image processing technology is adopted to obtain standard images and real-time images of contact lines, pre-process, crop, stitching and chunking processing are performed, and image matching is used to match with feature point, and the disease detection model and loss value calculation is combined to output wear detection results.
It improves the accuracy and efficiency of contact line wear detection, can promptly detect wear warnings, reduce manual workload, and reduce the impact of environmental factors.
Smart Images
Figure CN120580461A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of contact line detection technology, and in particular to a method and system for detecting wear of a rail transit contact line. Background Art
[0002] Railways are a vital component of modern transportation systems, and their safety and reliability are directly linked to economic development and property security. Within the railway system, the catenary system is a crucial component of the power supply system, providing electrical energy for electric locomotives. As the core of the catenary system, the condition of the contact wire directly impacts the quality of current collection and operational safety of the train. However, with frequent train operation and the passage of time, the contact wire experiences mechanical and electrical wear as it slides against the pantograph. If this wear reaches a certain level and is not addressed promptly, it not only degrades the conductive performance of the contact wire but can also cause power failures and even disrupt the normal operation of the train.
[0003] Traditional contact wire wear detection relies primarily on manual inspections and mechanical measuring instruments. While manual inspections are intuitive, manual wear measurement is labor-intensive, prone to missed inspections, and inefficient. Furthermore, it is difficult to provide a comprehensive and accurate assessment of contact wire wear. While mechanical measuring instruments can quantitatively measure contact wire wear, they are complex to operate and susceptible to environmental factors, leading to errors in the measurement results. Therefore, finding an efficient and accurate contact wire wear detection method is crucial. Summary of the Invention
[0004] In view of this, the present application provides a method and system for detecting contact line wear in rail transit to address the deficiencies in the prior art.
[0005] In a first aspect, the present application provides a method for detecting wear of a rail transit contact wire, comprising: Obtain a standard image of the contact line to be measured and place it in a standard image set; Use the preset equipment to collect the real-time images of the contact line to be tested, and put all the real-time images into the real-time image set after pre-processing; Based on the anchor segments, the real-time images collected in the real-time image set are cropped and spliced to obtain collected images of contact lines between multiple anchor segments; Based on the equidistant segmentation processing of the anchor segment, the corresponding acquired image is divided into blocks to obtain multiple block images; All the block images of the contact line between any anchor segments are matched in the standard image set based on the feature point matching method, and multiple matching results are output; All matching results are processed using a preset method, and the wear detection results of the contact lines between any anchor segments are output.
[0006] In a possible implementation of the first aspect, preprocessing all real-time images includes: performing denoising and correction processing on all real-time images.
[0007] In a possible implementation of the first aspect, based on the anchor segments, cropping and splicing the real-time images collected in the real-time image set to obtain collected images of contact lines between multiple anchor segments includes: Get the two endpoints of any anchor segment and record them as point A and point B respectively; Acquire a real-time image including point A as a first real-time image, and acquire a real-time image including point B as a second real-time image; Cropping the first real-time image with point A as the boundary, and obtaining a plurality of feature points on the edge away from point A as first feature points; Based on all the first feature points, obtaining an image adjacent to the first real-time image in the real-time image set and recording it as a third real-time image; The third real-time image is cropped with all the first feature points as boundaries, and a plurality of feature points are obtained at edges away from all the first feature points and recorded as second feature points; Based on all the second feature points, adjacent images of the third real-time image are acquired in the real-time image set until the second real-time image, and the second real-time image is cropped with point B as the boundary; The cropped first real-time image and the cropped second real-time image, as well as all images therebetween, are sequentially spliced to obtain an acquired image of the contact line between any anchor segments.
[0008] In a possible implementation of the first aspect, all the block images are matched in the standard image set based on feature point matching, and outputting multiple matching results includes: Setting a plurality of feature points in each block image of the contact line between any anchor segments, matching the feature points with the corresponding images in the standard image set, and outputting the contact line disease type and contact line loss value corresponding to each block image; The contact line disease types and contact line loss values corresponding to all block images together constitute the matching result.
[0009] In a possible implementation of the first aspect, outputting the contact line defect type corresponding to each block image includes: Constructing a disease detection model and training the disease detection model using labeled contact line disease images; Annotating residual regions in a single block image after matching with a corresponding image in the standard image set; The block image containing the annotated residual area is input into the trained defect detection model to output the corresponding contact line defect type.
[0010] In a possible implementation of the first aspect, outputting the contact line loss value corresponding to each block image includes: Obtaining the contour of the contact line in a single block image, and performing subtraction processing on the contour containing the contact line in the corresponding image in the standard image set; The average contour value of the regional contact line obtained by difference processing is calculated as the corresponding contact line loss value.
[0011] In a possible implementation of the first aspect, the preset method includes: Obtain any anchor segment as a first anchor segment, and record all matching results of contact lines between the first anchor segments as first matching results; Based on all the first matching results, obtaining the contact line damage type corresponding to each block image, and outputting it as the damage type of each contact line segment between the first anchor segments; Based on all first matching results, obtaining a contact line loss value corresponding to each block image and using the contact line loss value of each contact line between the first anchor segments; Based on the loss value of each contact line between the first anchor segments, a first judgment is performed to obtain a first judgment result; if a warning of wear of the contact line between the first anchor segments is output based on the first judgment result, no action is taken; otherwise, a second judgment is performed to obtain a second judgment result; Based on the second judgment result, determining whether there is a wear warning for the contact line between the first anchor segments, and if so, outputting a wear warning for the contact line between the first anchor segments; if not, outputting a wear warning for the contact line between the first anchor segments; The wear detection result of the contact line between the first anchor sections includes the damage type of each contact line section and whether there is a wear warning.
[0012] In a possible implementation of the first aspect, the first-level determination includes: The number of segments between the first anchor segments during exhalation that pass through the segmented contact line is recorded as N segments; The number of segments corresponding to the loss value of the single-segment contact line between the first anchor segments exceeding the threshold is recorded as M segments; Determine whether the ratio of M / N exceeds a first set value, and if so, output a warning indicating wear of the contact line between the first anchor segments; if not, take no action; Secondary judgments include: Acquire contact lines whose loss values of single-segment contact lines between the first anchor segments continuously exceed a threshold, and store them in a preset data set in a grouped manner; Determine whether the preset data set contains a single group of contact line segments with a number exceeding a second set value, and if so, clean the contact line data of the remaining groups; Based on the preset data set after cleaning, it is determined whether there are multiple contact lines in a single group of contact lines, and the loss values are increased in sequence. If so, a warning of wear of the contact lines between the first anchor segments is output; if not, a warning of no wear of the contact lines between the first anchor segments is output.
[0013] In a possible implementation manner of the first aspect, the preset device is a high-speed industrial camera.
[0014] A second aspect of the present application provides a rail transit contact wire wear detection system, comprising: The marking image unit is used to obtain the standard image of the contact line to be measured and put it into the standard image set; A real-time image unit is used to collect real-time images of the contact line to be tested using a preset device, and to pre-process all real-time images and then place them into a real-time image set; An image acquisition unit, configured to crop and splice the real-time images acquired from the real-time image set based on the anchor segments to obtain acquired images of contact lines between multiple anchor segments; A blocking unit is used to block the corresponding acquired image to obtain a plurality of block images based on the equidistant segmentation processing of the anchor segment; A matching unit, configured to perform image matching on all block images of the contact line between any anchor segments in the standard image set based on feature point matching, and output a plurality of matching results; The wear detection unit is used to process all matching results using a preset method and output the wear detection results of the contact lines between any anchor segments.
[0015] Its beneficial effects are: The present invention discloses a method and system for detecting wear of rail transit contact lines, which obtains a standard image of the contact line to be measured and puts it into a standard image set; uses a real-time image of the contact line to be measured through a preset device, and pre-processes the real-time image including denoising and correction; takes the anchor segment as a reference, crops and splices the acquired real-time image to obtain a captured image of the contact line between each anchor segment; takes the equidistant segmentation of the anchor segment as a reference, blocks the captured image between a single anchor segment to obtain multiple block images; then, based on the feature point matching method, matches all the block images between a single anchor segment with the corresponding images in the standard image set, and outputs multiple matching results; then, analyzes and processes all the matching results, integrates them and outputs the wear detection results of the contact line between a single anchor segment. The present invention collects real-time images of contact lines, and based on a single anchor segment, crops and splices the real-time images of contact lines between single anchor segments to obtain collected images of contact lines between single anchor segments. The contact lines between single anchor segments are then segmented, that is, the collected images of contact lines between single anchor segments are divided into blocks to obtain multiple block images. The block images are matched with standard images in a standard image set based on feature points to output multiple matching results. The block processing and separate comparison methods can not only detect diseases with slight corresponding features of the disease in the contact line (improving the detection accuracy), but also improve the detection efficiency. Finally, all matching results are analyzed and integrated to output the wear detection results of the contact lines between single anchor segments. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without any creative work.
[0017] Figure 1 This is a flow chart of a method for detecting contact line wear in rail transit provided in an embodiment of the present application; Figure 2 This is a schematic diagram of the composition of a rail transit contact line wear detection system provided in an embodiment of the present application. DETAILED DESCRIPTION
[0018] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.
[0019] In this application, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus comprising the element.
[0020] In order to better understand this application, the following are corresponding explanations of the technical names involved in this application: Example 1 In existing technologies, contact wire wear detection primarily relies on manual inspections and mechanical measuring instruments. While manual inspections are intuitive, manual wear measurement is labor-intensive, prone to missed inspections, and inefficient. Furthermore, it is difficult to provide a comprehensive and accurate assessment of contact wire wear. While mechanical measuring instruments can quantitatively measure contact wire wear, they are complex to operate and are susceptible to environmental factors, leading to measurement errors.
[0021] Therefore, the present application provides a method for detecting wear of contact wires in rail transit, such as Figure 1 Shown, including: Obtain a standard image of the contact line to be measured and place it in a standard image set; Use the preset equipment to collect the real-time images of the contact line to be tested, and put all the real-time images into the real-time image set after pre-processing; Based on the anchor segments, the real-time images collected in the real-time image set are cropped and spliced to obtain collected images of contact lines between multiple anchor segments; Based on the equidistant segmentation processing of the anchor segment, the corresponding acquired image is divided into blocks to obtain multiple block images; All the block images of the contact line between any anchor segments are matched in the standard image set based on the feature point matching method, and multiple matching results are output; All matching results are processed using a preset method, and the wear detection results of the contact lines between any anchor segments are output.
[0022] Among them, the standard image of the contact line to be measured is obtained and placed in the standard image set; the standard image of the contact line to be measured is obtained through the standard image library of the contact line.
[0023] Among them, a preset device is used to collect real-time images of the contact line to be tested. The preset device adopts a high-speed industrial camera, and all real-time images are denoised and corrected, and then placed in a real-time image set.
[0024] The real-time images collected in the real-time image set are cropped and spliced based on the anchor segments to obtain images of the contact lines between multiple anchor segments. Specifically, the two endpoints of a single contact line between anchor segments are first found in the real-time image set, denoted as points A and B. A real-time image containing point A and a real-time image containing point B are then acquired. Furthermore, real-time images between the real-time image containing point A and the real-time image containing point B are acquired. All acquired real-time images are cropped and then spliced together. For real-time images containing both points A and B, cropping is performed using point A or B as the boundary. For all real-time images between the real-time image containing point A and the real-time image containing point B, splicing is performed sequentially based on matching boundary feature points.
[0025] Among them, based on the equidistant segmentation processing of the anchor segments, the corresponding acquired images are segmented to obtain multiple block images; if the contact line between single anchor segments is equidistantly segmented into T segments, the corresponding acquired images should also be segmented to obtain T block images.
[0026] Among them, all block images of the contact line between any anchor segments are matched in the standard image set based on the feature point matching method, and multiple matching results are output; if there are T block images for the contact line between a single anchor segment, then multiple feature points are set in the T block images respectively, and matched with the standard image in the standard image set based on the feature point matching method, firstly, the residual area corresponding to each block image is output, and all residual areas are identified by constructing a disease detection model, and finally the corresponding disease type is output; then the difference between the contour of the contact line corresponding to each block image and the contour of the contact line corresponding to the standard image is output, which is the loss value of the contact line, that is, the matching result includes the disease type of the contact line and the loss value of the contact line.
[0027] Among them, a preset method is used to process all matching results and output the wear detection results of the contact line between any anchor segments; the preset method is specifically as follows: first, all matching results of the contact line between any anchor segments are obtained, recorded as the first matching result, and the disease type of the contact line is directly output through all the first matching results; for the loss value of the contact line, primary and secondary judgments are performed, and the primary judgment is to first calculate the number of segments (such as M) whose loss value of the contact line between a single anchor segment exceeds the threshold, and the total number of segments of the contact line between a single anchor segment is N, and judge whether the ratio of M / N exceeds the first set value. If so, it means that there are many places where the contact line between the anchor segments is severely worn. At this time, there is a greater risk of wear on the contact line, so it is judged that there is a wear warning; if not, a secondary judgment is performed, and the secondary judgment is to first obtain the contact lines whose loss values between the single anchor segments continuously exceed the threshold, and group them and store them in the preset data set. For example, there are 5 groups (M1 , M3, M4 and M5) the loss value of each contact line segment exceeds the threshold, and the contact lines in each group are continuous, the number of contact line segments in M1 is 2 segments, the number of contact line segments in M2 is 3 segments, the number of contact line segments in M3 is 4 segments, the number of contact line segments in M4 is 2 segments, and the number of contact line segments in M5 is 5 segments. Then, it is determined whether the number of contact line segments in these 5 groups exceeds the second set value (if it is 3). If so, the M3 group and the M5 group are retained, and then it is determined whether the loss values of adjacent contact lines in the M3 group and the M5 group are continuously increased. For example, the loss values of the continuous contact lines in the M3 group are a1>a2<a3>a4, that is, the corresponding change of the contact line is from thin to thick, then thinner, and then thicker. The loss values of the continuous contact lines in the M5 group are b1<b2<b3<b4<b5, that is, the contact line is gradually becoming thinner. At this time, there is a greater risk of wear of the contact line, so a wear warning of the contact line is output. Otherwise, a wear warning of the contact line is output.
[0028] In some embodiments, pre-processing all real-time images includes: performing denoising and correction processing on all real-time images.
[0029] In some embodiments, based on the anchor segments, cropping and splicing the real-time images collected in the real-time image set to obtain collected images of contact lines between multiple anchor segments includes: Get the two endpoints of any anchor segment and record them as point A and point B respectively; Acquire a real-time image including point A as a first real-time image, and acquire a real-time image including point B as a second real-time image; Cropping the first real-time image with point A as the boundary, and obtaining a plurality of feature points on the edge away from point A as first feature points; Based on all the first feature points, obtaining an image adjacent to the first real-time image in the real-time image set and recording it as a third real-time image; The third real-time image is cropped with all the first feature points as boundaries, and a plurality of feature points are obtained at edges away from all the first feature points and recorded as second feature points; Based on all the second feature points, adjacent images of the third real-time image are acquired in the real-time image set until the second real-time image, and the second real-time image is cropped with point B as the boundary; The cropped first real-time image and the cropped second real-time image, as well as all images therebetween, are sequentially spliced to obtain an acquired image of the contact line between any anchor segments.
[0030] In some embodiments, all the block images are matched in the standard image set based on feature point matching, and the output of multiple matching results includes: Setting a plurality of feature points in each block image of the contact line between any anchor segments, matching the feature points with the corresponding images in the standard image set, and outputting the contact line disease type and contact line loss value corresponding to each block image; The contact line disease types and contact line loss values corresponding to all block images together constitute the matching result.
[0031] In some embodiments, outputting the contact line damage type corresponding to each block image includes: Constructing a disease detection model and training the disease detection model using labeled contact line disease images; Annotating residual regions in a single block image after matching with a corresponding image in the standard image set; The block image containing the annotated residual area is input into the trained defect detection model to output the corresponding contact line defect type.
[0032] In some embodiments, outputting the contact line loss value corresponding to each block image includes: Obtaining the contour of the contact line in a single block image, and performing subtraction processing on the contour containing the contact line in the corresponding image in the standard image set; The average contour value of the regional contact line obtained by difference processing is calculated as the corresponding contact line loss value.
[0033] In some embodiments, the preset method includes: Obtain any anchor segment as a first anchor segment, and record all matching results of contact lines between the first anchor segments as first matching results; Based on all the first matching results, obtaining the contact line damage type corresponding to each block image, and outputting it as the damage type of each contact line segment between the first anchor segments; Based on all first matching results, obtaining a contact line loss value corresponding to each block image and using the contact line loss value of each contact line between the first anchor segments; Based on the loss value of each contact line between the first anchor segments, a first judgment is performed to obtain a first judgment result; if a warning of wear of the contact line between the first anchor segments is output based on the first judgment result, no action is taken; otherwise, a second judgment is performed to obtain a second judgment result; Based on the second judgment result, determining whether there is a wear warning for the contact line between the first anchor segments, and if so, outputting a wear warning for the contact line between the first anchor segments; if not, outputting a wear warning for the contact line between the first anchor segments; The wear detection result of the contact line between the first anchor sections includes the damage type of each contact line section and whether there is a wear warning.
[0034] In some embodiments, the first-level determination includes: The number of segments between the first anchor segments during exhalation that pass through the segmented contact line is recorded as N segments; The number of segments corresponding to the loss value of the single-segment contact line between the first anchor segments exceeding the threshold is recorded as M segments; Determine whether the ratio of M / N exceeds a first set value, and if so, output a warning indicating wear of the contact line between the first anchor segments; if not, take no action; Secondary judgments include: Acquire contact lines whose loss values of single-segment contact lines between the first anchor segments continuously exceed a threshold, and store them in a preset data set in a grouped manner; Determine whether the preset data set contains a single group of contact line segments with a number exceeding a second set value, and if so, clean the contact line data of the remaining groups; Based on the preset data set after cleaning, it is determined whether there are multiple contact lines in a single group of contact lines, and the loss values are increased in sequence. If so, a warning of wear of the contact lines between the first anchor segments is output; if not, a warning of no wear of the contact lines between the first anchor segments is output.
[0035] In some embodiments, the preset device is a high-speed industrial camera.
[0036] Example 2 Based on the method for detecting wear of a contact line of rail transit provided in the first embodiment of the present application, the second embodiment of the present application also provides a system for detecting wear of a contact line of rail transit, such as Figure 2 As shown, the system includes: The marking image unit is used to obtain the standard image of the contact line to be measured and put it into the standard image set; A real-time image unit is used to collect real-time images of the contact line to be tested using a preset device, and to pre-process all real-time images and then place them into a real-time image set; An image acquisition unit, configured to crop and splice the real-time images acquired from the real-time image set based on the anchor segments to obtain acquired images of contact lines between multiple anchor segments; A blocking unit is used to block the corresponding acquired image to obtain a plurality of block images based on the equidistant segmentation processing of the anchor segment; A matching unit, configured to perform image matching on all block images of the contact line between any anchor segments in the standard image set based on feature point matching, and output a plurality of matching results; The wear detection unit is used to process all matching results using a preset method and output the wear detection results of the contact lines between any anchor segments.
[0037] The specific principles and execution processes of each unit in the rail transit contact line wear detection system disclosed in the above-mentioned embodiment 2 of the present application are the same as the rail transit contact line wear detection method disclosed in the above-mentioned embodiment 1 of the present application. Please refer to the corresponding parts of the rail transit contact line wear detection method disclosed in the above-mentioned embodiment 1 of the present application, and no further details will be given here.
[0038] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computing software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0039] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0040] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.
Claims
1. A method for detecting wear of contact wires in rail transit, characterized in that: include: Obtain a standard image of the contact line to be measured and place it in a standard image set; Use the preset equipment to collect the real-time images of the contact line to be tested, and put all the real-time images into the real-time image set after pre-processing; Based on the anchor segments, the real-time images collected in the real-time image set are cropped and spliced to obtain collected images of contact lines between multiple anchor segments; Based on the equidistant segmentation processing of the anchor segment, the corresponding acquired image is divided into blocks to obtain multiple block images; All the block images of the contact line between any anchor segments are matched in the standard image set based on the feature point matching method, and multiple matching results are output; All matching results are processed using a preset method, and the wear detection results of the contact lines between any anchor segments are output.
2. A method for detecting contact line wear in rail transit according to claim 1, characterized in that: Preprocessing all real-time images includes: performing denoising and correction processing on all real-time images.
3. The method for detecting contact line wear in rail transit according to claim 1, characterized in that: Based on the anchor segments, the real-time images collected in the real-time image set are cropped and spliced to obtain collected images of contact lines between multiple anchor segments, including: Get the two endpoints of any anchor segment and record them as point A and point B respectively; Acquire a real-time image including point A as a first real-time image, and acquire a real-time image including point B as a second real-time image; Cropping the first real-time image with point A as the boundary, and obtaining a plurality of feature points on the edge away from point A as first feature points; Based on all the first feature points, obtaining an image adjacent to the first real-time image in the real-time image set and recording it as a third real-time image; The third real-time image is cropped with all the first feature points as boundaries, and a plurality of feature points are obtained at edges away from all the first feature points and recorded as second feature points; Based on all the second feature points, adjacent images of the third real-time image are acquired in the real-time image set until the second real-time image, and the second real-time image is cropped with point B as the boundary; The cropped first real-time image and the cropped second real-time image, as well as all images therebetween, are sequentially spliced to obtain an acquired image of the contact line between any anchor segments.
4. The method for detecting contact line wear in rail transit according to claim 1, characterized in that: All block images are matched in the standard image set based on feature point matching, and multiple matching results are output including: Setting a plurality of feature points in each block image of the contact line between any anchor segments, matching the feature points with the corresponding images in the standard image set, and outputting the contact line disease type and contact line loss value corresponding to each block image; The contact line disease types and contact line loss values corresponding to all block images together constitute the matching result.
5. The method for detecting contact line wear in rail transit according to claim 4, characterized in that: The contact line damage types corresponding to each block image output include: Constructing a disease detection model and training the disease detection model using labeled contact line disease images; Annotating residual regions in a single block image after matching with a corresponding image in the standard image set; The block image containing the annotated residual area is input into the trained defect detection model to output the corresponding contact line defect type.
6. A method for detecting contact line wear in rail transit according to claim 4, characterized in that: The contact line loss values corresponding to each block image are output as follows: Obtaining the contour of the contact line in a single block image, and performing subtraction processing on the contour containing the contact line in the corresponding image in the standard image set; The average contour value of the regional contact line obtained by difference processing is calculated as the corresponding contact line loss value.
7. A method for detecting contact line wear in rail transit according to claim 4, characterized in that: The preset method includes: Obtain any anchor segment as a first anchor segment, and record all matching results of contact lines between the first anchor segments as first matching results; Based on all the first matching results, obtaining the contact line damage type corresponding to each block image, and outputting it as the damage type of each contact line segment between the first anchor segments; Based on all first matching results, obtaining a contact line loss value corresponding to each block image and using the contact line loss value of each contact line between the first anchor segments; Based on the loss value of each contact line between the first anchor segments, a first judgment is performed to obtain a first judgment result; if a warning of wear of the contact line between the first anchor segments is output based on the first judgment result, no action is taken; otherwise, a second judgment is performed to obtain a second judgment result; Based on the second judgment result, determining whether there is a wear warning for the contact line between the first anchor segments, and if so, outputting a wear warning for the contact line between the first anchor segments; if not, outputting a wear warning for the contact line between the first anchor segments; The wear detection result of the contact line between the first anchor sections includes the damage type of each contact line section and whether there is a wear warning.
8. A method for detecting contact line wear in rail transit according to claim 7, characterized in that: The first-level judgment includes: The number of segments between the first anchor segments during exhalation that pass through the segmented contact line is recorded as N segments; The number of segments corresponding to the loss value of the single-segment contact line between the first anchor segments exceeding the threshold is recorded as M segments; Determine whether the ratio of M / N exceeds a first set value, and if so, output a warning indicating wear of the contact line between the first anchor segments; if not, take no action; Secondary judgments include: Acquire contact lines whose loss values of single-segment contact lines between the first anchor segments continuously exceed a threshold, and store them in a preset data set in a grouped manner; Determine whether the preset data set contains a single group of contact line segments with a number exceeding a second set value, and if so, clean the contact line data of the remaining groups; Based on the preset data set after cleaning, it is determined whether there are multiple contact lines in a single group of contact lines, and the loss values are increased in sequence. If so, a warning of wear of the contact lines between the first anchor segments is output; if not, a warning of no wear of the contact lines between the first anchor segments is output.
9. The method for detecting contact line wear in rail transit according to claim 1, characterized in that: The preset device is a high-speed industrial camera.
10. A contact line wear detection system for rail transit, characterized in that: include: The marking image unit is used to obtain the standard image of the contact line to be measured and put it into the standard image set; A real-time image unit is used to collect real-time images of the contact line to be tested using a preset device, and to pre-process all real-time images and then place them into a real-time image set; An image acquisition unit, configured to crop and splice the real-time images acquired from the real-time image set based on the anchor segments to obtain acquired images of contact lines between multiple anchor segments; A blocking unit is used to block the corresponding acquired image to obtain a plurality of block images based on the equidistant segmentation processing of the anchor segment; A matching unit, configured to perform image matching on all block images of the contact line between any anchor segments in the standard image set based on feature point matching, and output a plurality of matching results; The wear detection unit is used to process all matching results using a preset method and output the wear detection results of the contact lines between any anchor segments.