A contact network limit stop nail surplus state detection system

By designing the contact network limit nail stop margin state detection system, using a combination of high-pixel and low-pixel deep learning detection networks, intelligent and automated detection of the contact network limit nail stop margin is achieved, and the existing technology is solved The lack of detection methods improves the accuracy and speed of analysis and reduces labor costs.

CN111754508BActive Publication Date: 2024-11-08YI TAI FEI LIU INFORMATION TECH LLC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010633369.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-07-06
Publication Date
2024-11-08
Estimated Expiration
2040-07-06

AI Technical Summary

Technical Problem

The existing technology lacks effective intelligent methods to detect the remaining state of the contact network limit positioner nail stop, which leads to time-consuming and labor-intensive inspection and easy to miss inspection. There is little research at home and abroad, and high-precision image detection methods are also used. Not widely used.

Method used

A contact network limit nail stop margin detection system is designed, including image acquisition module, image positioning module, nail stop margin calculation module and analysis module. It uses a high-pixel and low-pixel deep learning detection network to pass edge pictures and the detection box calculates and extracts the nail stop margin to achieve automated and intelligent analysis.

Benefits of technology

Improve the accuracy and speed of the state analysis of nail stop margin, reduce labor costs, avoid manual inspection errors, realize intelligent and automated detection of the contact network limit nail stop margin, and solve the existing technology Inadequate detection methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111754508B_ABST
    Figure CN111754508B_ABST
Patent Text Reader

Abstract

The present invention discloses a contact network limit stop nail margin state detection system, comprising an image acquisition module, an image positioning module, a stop nail margin calculation module, a stop nail margin analysis module and an analysis result output module, wherein the image positioning module comprises a high-pixel locator detection network and a low-pixel stop nail detection network, and the present invention relates to the contact network limit stop nail margin state detection technical field. The contact network limit stop nail margin state detection system intelligently analyzes the stop nail margin state in the high-speed railway contact network image, and improves the analysis speed while increasing the accuracy of the margin state analysis, and compared with the existing method of manually checking the stop nail margin state, the present invention uses the contact network high-definition image to automatically and intelligently analyze whether the contact network limit stop nail margin of the limit locator is too much or too little, thereby reducing the labor cost and avoiding possible errors in manual checking.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of contact network limit stop nail surplus state detection, and in particular to a contact network limit stop nail surplus state detection system. Background Art

[0002] The overhead contact network is an important equipment in the power supply system of high-speed railways. It has the characteristics of complex structure and open-air setting. Once a problem occurs, it is very likely to affect the power supply system of the entire railway and the safety of the running trains. The overhead contact network suspension status detection and monitoring system (4C) can eliminate potential faults in time and ensure the good operation of the components by performing high-precision imaging of the railway overhead contact network components.

[0003] The margin of the limit stopper of the limit locator is an important parameter in the contact network components. When the limit locator is used, the locator and the stopper should have an appropriate margin. However, due to long-term sun and rain, mechanical stress, strong electromagnetic fields and even vibrations caused by train running, the stopper is very likely to loosen the screws, resulting in changes in the margin, which may even affect the normal operation of the train and cause accidents in serious cases.

[0004] The existing method for analyzing the residual state of the limit stop pins of the contact network limit locator mainly adopts a combination of manual regular inspection and technical personnel checking the 4C contact network high-precision images. The former has high labor intensity, long working cycle, and it is not easy to find defects. As for the latter, although the workload of investigation is greatly reduced, the technicians still have to find the limit locator from the huge amount of high-precision image data of the contact network, and then analyze and judge the residual state of the limit stop pin, which is time-consuming and labor-intensive, and it is very easy to miss the detection due to visual fatigue. Therefore, designing a system for automatic and intelligent detection and monitoring of various faults for high-precision images of the contact network has become a problem that needs to be faced. Designing an intelligent system for detecting the residual state of the limit locator stop pin is of great practical significance for ensuring the safe operation of the contact network.

[0005] At present, there is no good method to detect whether there are too many or insufficient limit locator stop nails in the high-definition images of the contact network. It is still at the stage of manually viewing high-definition images. There is little research on intelligent analysis of the limit locator stop nail remainder status of the contact network at home and abroad, and there is almost no method to use high-precision images to detect the limit locator stop nail remainder status. Summary of the invention

[0006] 1. Technical issues to be resolved

[0007] In view of the deficiencies in the prior art, the present invention provides a contact network limit stop nail remainder status detection system, which solves the current problem that there is no good method for detecting whether there are two defect states of excessive or insufficient limit locator stop nail remainder in the contact network high-definition image, and the method still remains in the stage of manually viewing high-definition pictures. There is very little research on the intelligent analysis of the contact network limit locator stop nail remainder status at home and abroad, and there is almost no method for detecting the limit locator stop nail remainder status using high-precision images.

[0008] (II) Technical solution

[0009] To achieve the above objectives, the present invention is implemented through the following technical solutions: a contact network limit stop nail margin status detection system, including an image acquisition module, an image positioning module, a stop nail margin calculation module, a stop nail margin analysis module and an analysis result output module, the image positioning module includes a high-pixel locator detection network and a low-pixel stop nail detection network, and the output end of the high-pixel locator detection network is connected to the input end of the low-pixel stop nail detection network.

[0010] Preferably, the nail stopping margin calculation module includes an edge image, a new detection frame, new nail stopping coordinates and a nail stopping margin.

[0011] Preferably, the output end of the edge image is connected to the input end of the new detection frame, and the output end of the edge image is connected to the input end of the new stop pin coordinates.

[0012] Preferably, the output end of the new detection frame is connected to the input end of the new nail-stopping coordinates, and the output end of the new nail-stopping coordinates is connected to the input end of the nail-stopping margin.

[0013] Preferably, the stop nail margin analysis module includes detection results and stop nail margin, positioning detection module and judgment result output.

[0014] Preferably, the output end of the detection result and the nail-stopping margin is connected to the input end of the positioning detection module, and the output end of the positioning detection module is connected to the input end of the judgment result output.

[0015] Preferably, the output end of the image acquisition module is connected to the input ends of the image positioning module, the nail stopping margin calculation module and the nail stopping margin analysis module respectively.

[0016] Preferably, the output ends of the image positioning module, the nail stopping margin calculation module and the nail stopping margin analysis module are all connected to the input end of the analysis result output module.

[0017] (III) Beneficial effects

[0018] The present invention provides a contact network limit stop nail surplus state detection system. Compared with the prior art, it has the following beneficial effects:

[0019] (1) The contact network limit stop nail remainder state detection system comprises an image acquisition module, an image positioning module, a stop nail remainder calculation module, a stop nail remainder analysis module and an analysis result output module, wherein the output end of the image acquisition module is respectively connected to the input end of the image positioning module, the stop nail remainder calculation module and the stop nail remainder analysis module, and the output ends of the image positioning module, the stop nail remainder calculation module and the stop nail remainder analysis module are all connected to the input end of the analysis result output module. By intelligently analyzing the stop nail remainder state in the high-speed railway contact network image, while increasing the accuracy of the remainder state analysis, the speed of the analysis is improved. Compared with the existing method of manually checking the stop nail remainder state, the present invention uses high-definition contact network images to automatically and intelligently analyze whether the contact network limit stop nail remainder is too much or too little, thereby greatly reducing the labor cost and avoiding the errors that may occur during manual checking.

[0020] (2) The contact network limit stop nail remainder state detection system includes a high-pixel locator detection network and a low-pixel stop nail detection network in the image positioning module, the output end of the high-pixel locator detection network is connected to the input end of the low-pixel stop nail detection network, the stop nail remainder calculation module includes an edge image, a new detection frame, a new stop nail coordinate and a stop nail remainder, the output end of the edge image is connected to the input end of the new detection frame, the output end of the edge image is connected to the input end of the new stop nail coordinate, the output end of the new detection frame is connected to the input end of the new stop nail coordinate, and the output end of the new stop nail coordinate is connected to the input end of the stop nail remainder. By adopting two deep learning detection models for high pixels and low pixels, the locator and the stop nail are identified respectively, so that the required detection frame can be accurately extracted. By extracting the edge of the image and calculating in coordination with the detection frame, the pixel value of the stop nail remainder is abstractly calculated, thereby indirectly bypassing the step of estimating the actual length of the stop nail remainder.

[0021] (3) The contact network limit stop nail remainder status detection system includes a stop nail remainder analysis module including a detection result and a stop nail remainder, a positioning detection module and a judgment result output, the output end of the detection result and the stop nail remainder is connected to the input end of the positioning detection module, and the output end of the positioning detection module is connected to the input end of the judgment result output. By abstractly converting the sufficiency of the stop nail remainder status into a method of calculating the ratio of the stop nail remainder pixel value to the stop nail length pixel, the stop nail remainder status can be well analyzed, thereby bypassing the problem that the locator size cannot be accurately quantified into a specific value due to the distance of the camera when the picture is taken. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A system structure provided for an embodiment of the present invention;

[0023] Figure 2 A flow chart of an image positioning method provided by an embodiment of the present invention;

[0024] Figure 3 A flow chart of a method for calculating a nail stop allowance provided by an embodiment of the present invention;

[0025] Figure 4 A flow chart of a method for analyzing the stop nail allowance provided in an embodiment of the present invention.

[0026] In the figure, 101, image acquisition module; 102, image positioning module; 201, high-pixel locator detection network; 202, low-pixel nail stopping detection network; 103, nail stopping margin calculation module; 301, edge image; 302, new detection frame; 303, new nail stopping coordinates; 304, nail stopping margin; 104, nail stopping margin analysis module; 401, detection result and nail stopping margin; 402, positioning detection module; 403, judgment result output; 105, analysis result output module. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] See also Figure 1 An embodiment of the present invention provides a technical solution: a contact network limit stop nail margin state detection system, including an image acquisition module 101, an image positioning module 102, a stop nail margin calculation module 103, a stop nail margin analysis module 104 and an analysis result output module 105.

[0029] Image acquisition module 101: an image acquisition module, which acquires high-precision contact network images from the contact network suspension state detection and monitoring system (4C);

[0030] Image positioning module 102: The image positioning module first inputs the contact network image output by module 101 into the deep learning detection model for high pixels, detects the detection frame of the limit locator, and crops the locator image, and then inputs the locator image into the deep learning detection model for low pixels to detect the detection frame of the stop nail and the stop nail ring;

[0031] The nail stop margin calculation module 103: The nail stop margin calculation module grayscales the locator image output by the 102 module and uses the double-threshold Canny edge detection operator to calculate the edge of the cropped limit locator image, and then combines the nail stop and the nail stop ring detection frame to obtain the nail stop margin;

[0032] The nail stop margin analysis module 104: The nail stop margin analysis module combines and calculates the nail stop margin output by the 103 module and the nail stop detection frame output by the 102 module, and analyzes whether the nail stop margin is within a preset threshold according to the ratio of the nail stop margin to the nail stop length, thereby obtaining an analysis result;

[0033] Analysis result output module 105: The analysis result output module marks the stop nail margin analysis result on the contact network image and outputs the final analysis result. It performs intelligent analysis on the stop nail margin status in the high-speed railway contact network image and increases the accuracy of margin status analysis while improving the speed of analysis. Compared with the existing method of manually checking the stop nail margin status, the present invention uses high-definition images of the contact network to automatically and intelligently analyze whether the stop nail margin of the contact network limit locator is too much or too little, which greatly reduces labor costs and avoids problems that may occur during manual inspection.

[0034] like Figure 2 The figure is a flow chart of the image positioning method. The image positioning module 102 includes a high-pixel locator detection network 201 and a low-pixel nail-stopping detection network 202. The output end of the high-pixel locator detection network 201 is connected to the input end of the low-pixel nail-stopping detection network 202. The specific implementation modules are as follows:

[0035] High-pixel locator detection network 201: Module 201 first feeds the contact network image into the YOLOV3 target detection network with an input of 1024*1024, obtains the limit locator detection frame in the image, and then crops the limit locator image from the original image;

[0036] Low-pixel stop nail detection network 202: Module 202 sends the limit locator image cropped in module 201 to the YOLOV3 target detection network with an input of 512*512, obtains the stop nail and stop nail clang detection frames in the image, and simultaneously outputs the cropped limit locator image, stop nail detection frame and stop nail clang detection frame. By adopting two deep learning detection models for high pixels and low pixels, the locator and stop nail are identified respectively, and the required detection frame can be accurately extracted.

[0037] like Figure 3 The figure is a flow chart of the method for calculating the stop nail margin. The stop nail margin calculation module 103 includes an edge image 301, a new detection frame 302, a new stop nail coordinate 303 and a stop nail margin 304. The output end of the edge image 301 is connected to the input end of the new detection frame 302, the output end of the edge image 301 is connected to the input end of the new stop nail coordinate 303, the output end of the new detection frame 302 is connected to the input end of the new stop nail coordinate 303, and the output end of the new stop nail coordinate 303 is connected to the input end of the stop nail margin 304. The specific implementation modules are as follows:

[0038] Edge image 301: Module 301 first processes the obtained limit locator image, performs grayscale conversion, binarization, and edge extraction using the Canny edge detection operator in sequence, and obtains an edge image of the limit locator;

[0039] New detection frame 302: Module 302 first combines the obtained nail-stopping detection frame with the nail-stopping bell detection frame, taking the minimum x-coordinate and the minimum y-coordinate of the upper left corner of the nail-stopping bell detection frame to generate new coordinates (xmin2, ymin2), and the maximum x-coordinate and the maximum y-coordinate of the lower right corner to generate new coordinates (xmax3, ymax3). Use the two new coordinate points to generate a new detection frame that includes both the nail-stopping bell and the nail-stopping bell;

[0040] New stop pin coordinates 303: Module 303 first intercepts the new detection frame edge image on the processed limit locator edge image according to the new detection frame coordinates obtained by module 302, and then calculates the new detection frame of the stop pin and ...

[0041] Stop nail margin 304: Module 304 first starts from the stop nail position in the new edge image generated by module 303, finds the pixel point with a pixel value equal to 255 near the new stop nail detection frame and closest to the new stop nail detection frame, sets it as the starting point s1, and then passes through pixel point s1 to continue traversing the horizontal coordinate in the direction of the new stop nail detection frame, and finds the second pixel point s2 with a pixel value equal to 255. The horizontal coordinate distance between s1 and s2 is the pixel distance of the stop nail margin, and outputs the stop nail margin. By extracting the edge of the picture and calculating in coordination with the detection frame, the pixel value of the stop nail margin is abstractly calculated, indirectly bypassing the step of estimating the actual length of the stop nail margin.

[0042] like Figure 4 The figure is a flow chart of the method for analyzing the nail stop margin. The nail stop margin analysis module 104 includes a detection result and nail stop margin 401, a positioning detection module 402 and a judgment result output 403. The output end of the detection result and nail stop margin 401 is connected to the input end of the positioning detection module 402, and the output end of the positioning detection module 402 is connected to the input end of the judgment result output 403. The specific implementation modules are as follows:

[0043] Detection results and nail stop margin 401: Module 401 obtains the locator detection results, nail stop and nail stop bell detection results, and nail stop margin pixel values;

[0044] Positioning detection module 402: Module 402 analyzes whether the nail stop margin is sufficient. When the following three conditions are met at the same time, the margin is determined to be appropriate: (1) the input locator detection result exists; (2) the input nail stop ring and the nail stop detection result exist at the same time and each is unique; (3) the ratio of the calculated nail stop margin pixel value to the nail stop width pixel value is within a preset threshold range;

[0045] Judgment result output 403: Module 403 outputs the analysis result obtained by module 402. By abstractly converting the sufficiency of the stop nail margin state into a method of calculating the ratio of the pixel value of the stop nail margin to the pixel of the stop nail length, the state of the stop nail margin can be well analyzed, thereby bypassing the problem that the locator size cannot be accurately quantified into a specific value due to the distance of the camera when the picture is taken.

[0046] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0047] It should be noted that, in this article, relational terms such as first and second, etc. are only used 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 "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0048] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A contact network limit stop nail margin state detection system, comprising an image acquisition module (101), an image positioning module (102), a stop nail margin calculation module (103), a stop nail margin analysis module (104) and an analysis result output module (105), characterized in that: The image positioning module (102) comprises a high-pixel locator detection network (201) and a low-pixel nail-stopping detection network (202), wherein an output end of the high-pixel locator detection network (201) is connected to an input end of the low-pixel nail-stopping detection network (202); Image acquisition module (101): the image acquisition module acquires a high-precision contact network image from the contact network suspension state detection and monitoring system (4C); Image positioning module (102): The image positioning module first inputs the contact network image output by the image acquisition module (101) into a deep learning detection model for high pixels, detects the detection frame of the limit locator, and crops the locator image, and then inputs the locator image into a deep learning detection model for low pixels, and detects the detection frames of the stop nail and the stop nail clang; The nail stop margin calculation module (103) performs grayscale processing on the locator image output by the image positioning module (102) and uses a double-threshold Canny edge detection operator to calculate the edge of the cropped limit locator image, and then combines the nail stop and the nail stop ring detection frame to obtain the nail stop margin; A nail stopping margin analysis module (104): The nail stopping margin analysis module combines and calculates the nail stopping margin output by the nail stopping margin calculation module (103) and the nail stopping detection frame output by the image positioning module (102), and analyzes whether the nail stopping margin is within a preset threshold value based on the ratio of the nail stopping margin to the nail stopping length, thereby obtaining an analysis result; Analysis result output module (105): The analysis result output module marks the stop nail margin analysis result on the contact network picture and outputs the final analysis result.

2. A contact network limit stop nail surplus state detection system according to claim 1, characterized in that: The nail stop margin calculation module (103) includes an edge image (301), a new detection frame (302), new nail stop coordinates (303) and a nail stop margin (304); The output end of the edge image (301) is connected to the input end of the new detection frame (302), and the output end of the edge image (301) is connected to the input end of the new nail stop coordinates (303); The output end of the new detection frame (302) is connected to the input end of the new nail stopping coordinates (303), and the output end of the new nail stopping coordinates (303) is connected to the input end of the nail stopping margin (304); the specific implementation module is as follows: Edge image (301): The edge image (301) first processes the obtained limit locator image, and sequentially performs grayscale conversion, binarization, and Canny edge detection operator edge extraction to obtain the edge image of the limit locator; New detection frame (302): The new detection frame (302) first combines the obtained stop nail detection frame with the stop nail bell detection frame, takes the minimum value of the x coordinate and the minimum value of the y coordinate of the upper left corner of the stop nail and the stop nail bell detection frame to generate new coordinates (xmin2, ymin2), and the maximum value of the x coordinate and the maximum value of the y coordinate of the lower right corner point to generate new coordinates (xmax3, ymax3), and uses the two new coordinate points to generate a new detection frame that includes both the stop nail and the stop nail bell; New stop nail coordinates (303): The new stop nail coordinates (303) are first obtained on the processed limit locator edge image according to the new detection frame coordinates obtained from the new detection frame (302), and the new detection frame edge image is intercepted, and then the new detection frame of the stop nail and the new detection frame of the stop nail and the new detection frame of the stop nail are calculated through the stop nail detection frame and the stop nail clang detection frame generated in the new detection frame (302); Stop nail margin (304): The stop nail margin (304) first starts from the stop nail position in the new edge image generated by the new stop nail coordinates (303), finds the pixel point with a pixel value equal to 255 near the new stop nail detection frame and closest to the new stop nail detection frame, and sets it as the starting point s1. Then, through the pixel point s1, continue to traverse the horizontal coordinate in the direction of the new stop nail detection frame, and find the second pixel point s2 with a pixel value equal to 255. The horizontal coordinate distance between s1 and s2 is the pixel distance of the stop nail margin, and outputs the stop nail margin. By extracting the edge of the picture and calculating in coordination with the detection frame, the pixel value of the stop nail margin is abstractly calculated, indirectly bypassing the step of estimating the actual length of the stop nail margin.

3. A contact network limit stop nail surplus state detection system according to claim 1, characterized in that: The nail stop margin analysis module (104) comprises a detection result and nail stop margin (401), a positioning detection module (402) and a judgment result output (403). The output end of the detection result and nail stop margin (401) is connected to the input end of the positioning detection module (402), and the output end of the positioning detection module (402) is connected to the input end of the judgment result output (403). The specific implementation module is as follows: Detection results and nail stop margin (401): The nail stop margin (401) obtains the locator detection result, the nail stop and nail stop bell detection results, and the nail stop margin pixel value; Positioning detection module (402): The positioning detection module (402) analyzes whether the nail stop margin is sufficient. When the following three conditions are met at the same time, it is determined that the margin is appropriate: (1) the input locator detection result exists; (2) the input nail stop ring and the nail stop detection result exist at the same time and each is unique; (3) the ratio of the calculated nail stop margin pixel value to the nail stop width pixel value is within a preset threshold range; Judgment result output (403): The judgment result output (403) outputs the analysis result obtained by the positioning detection module (402).

4. A contact network limit stop nail surplus state detection system according to claim 1, characterized in that: The output end of the image acquisition module (101) is respectively connected to the input ends of the image positioning module (102), the nail stopping margin calculation module (103) and the nail stopping margin analysis module (104).

5. The contact network limit stop nail surplus state detection system according to claim 1 is characterized by: The output ends of the image positioning module (102), the nail stopping margin calculation module (103) and the nail stopping margin analysis module (104) are all connected to the input end of the analysis result output module (105).