A device for detecting wear of subway power supply contact network wire

Abstract

The embodiment of the present application provides a device for detecting wear of a subway power supply contact network line, comprising a detection module, the detection module including an infrared laser and an infrared camera, the infrared laser being used to emit an infrared beam to the surface to be tested of the contact network line and forming a detection area on the surface of the contact network line, the infrared camera being used to capture an image to be tested of the detection area; the infrared laser being rotatably arranged at the open end of a U-shaped frame by a rotating support shaft; when the infrared camera is used to capture an image of the detection area, the image capture position of the infrared camera is adjusted by a lateral displacement structure; and an image processor being further provided, the image processor being used to receive the image to be tested captured by the infrared camera, the image processor having a built-in image recognition module, the image recognition module being used to identify the wear condition of the detection area based on the image to be tested. The detection device provided by the embodiment of the present application effectively solves the technical problem of relatively low detection result accuracy of existing detection devices.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of contact network line wear detection, and specifically, to a device for detecting wear of a subway power supply contact network line. Background Art

[0002] With the rapid development of urban rail transit, the subway has become an important part of the modern urban transportation system.

[0003] In subway systems, the power catenary wire is a key component that provides power to trains. Its performance and reliability directly impact the safe and stable operation of the subway. As train speeds continue to increase and traction currents continue to rise, the inherently weak elasticity of the rigidly suspended catenary wire leads to a gradual deterioration of the pantograph-catenary relationship, which in turn causes a series of contact wire wear issues.

[0004] Specifically, the contact wire wear is accelerated, with uneven wear distribution and even wavy wear in some localized areas. These wear issues can lead to unstable contact between the catenary and the pantograph, resulting in sparks and unusual noises between the catenary and the pantograph. Furthermore, the carbon slide surface can show signs of scratches and the contact wire can be chipped, seriously affecting the stability of the pantograph-pantograph connection and posing a significant risk to subway operation safety.

[0005] For contact line wear detection, the current main method used is to measure discrete points manually with a handheld vernier caliper. This method has relatively low detection accuracy and cannot meet the needs of accurate detection and analysis of the wear status of the line contact network. Summary of the Invention

[0006] The purpose of the embodiments of the present application is to provide a device for detecting wear of subway power supply contact network lines, so as to at least solve the technical problem that the detection results of existing detection devices are relatively low in accuracy and cannot meet the needs of accurate detection and analysis of the wear status of the line contact network.

[0007] To achieve the above objectives, the embodiments of the present application provide the following technical solutions.

[0008] According to one embodiment of the present application, a device for detecting wear of a subway power supply contact network line is provided, comprising:

[0009] The detection module includes an infrared laser and an infrared camera. The infrared laser is used to emit an infrared beam to the surface to be tested of the contact network line and form a detection area on the surface of the contact network line. The infrared camera is used to take pictures of the detection area to collect the image to be tested. The infrared laser is rotatably arranged at the open end of the U-shaped frame by rotating the support shaft. The infrared laser is used to generate a circular area beam. The central axis of the circular area beam and the central axis of the contact network line are always in the same plane A. The distance between the infrared laser and the contact network line is adjustable. When using the infrared camera to collect images of the detection area, the image collection position of the infrared camera is adjusted by the lateral displacement structure. The lateral displacement structure makes the displacement direction axis of the infrared camera coplanar with plane A, and the displacement direction axis is parallel to the central axis of the contact network line.

[0010] The image processor is used to receive the image to be inspected captured by the infrared camera. The image processor has a built-in image recognition module, which is used to identify the wear condition of the inspection area based on the image to be inspected.

[0011] Preferably, the angle between the central axis of the circular area light beam and the central axis of the contact network line is 30°-60°.

[0012] Preferably, a first micro servo motor is provided on the open end of the U-shaped frame;

[0013] The output shaft of the first micro servo motor is connected to the rotating support shaft fixedly connected to the infrared laser through a coupling;

[0014] A limit slider is fixedly installed on the U-shaped frame, an electric telescopic rod is installed on the limit slider, and the telescopic end of the electric telescopic rod is fixedly connected to the inside of the bottom shell;

[0015] A guide rail is fixedly arranged inside the bottom shell, and the guide rail cooperates with the limit slider. The guide rail is used to guide the movement of the limit slider.

[0016] Preferably, the lateral displacement structure comprises:

[0017] A displacement slider, the infrared camera is mounted on the displacement slider, the displacement slider is sleeved on the adjustment screw by a threaded connection, and the rotation of the adjustment screw is driven by a second micro servo motor; the second micro servo motor is fixed on the mounting frame;

[0018] The guide slide rod is slidably arranged on the displacement slider. The guide slide rod and the adjustment screw rod are arranged parallel to each other to ensure that the displacement slider can only move but not rotate; one end of the guide slide rod is also fixedly mounted on the mounting frame.

[0019] Preferably, the bottom shell is fixedly arranged on the detection box shell, the upper end of the bottom shell is an open structure, and the opening at the upper end of the bottom shell is communicated with the bottom opening of the detection box shell;

[0020] The infrared laser and the infrared camera are arranged in the bottom shell;

[0021] A box cover is provided on one side of the detection box shell in an opening and closing manner through an opening and closing hinge;

[0022] The detection box shell and the box cover are both provided with limit clamps, and the limit clamps adopt a split structure.

[0023] Preferably, the image recognition module is used to implement the following steps:

[0024] S101, preprocessing an image to be inspected, wherein the image to be inspected is one or more images captured by an infrared camera;

[0025] S102. Using the Otsu algorithm, the pre-processed image is divided into a dark area and a background. After segmentation, the dark area pixels are marked as 1 and the background pixels are marked as 0, thereby obtaining a binary image. Based on morphological operations, a morphologically processed binary image is output to highlight continuous dark areas. The morphologically processed binary image is traversed to mark independent connected domains. The area, minimum bounding rectangle, and grayscale mean of each connected domain are calculated. Pixels with an area smaller than a threshold are determined to be noise and filtered. Pixels with a grayscale mean much lower than the background are identified as defective dark areas, and their coordinates and size are recorded as areas of minor wear.

[0026] S103. Based on the location of the small wear area, adjust the infrared laser so that the edge contour line of the detection area formed by the infrared laser on the contact network line is within the small wear area; verify the authenticity of the wear in the small wear area using the Canny edge contour detection algorithm, and determine the wear result of the small wear area based on the verification result;

[0027] S104: Based on the wear results of multiple small wear areas, the wear condition of the current section of the contact network line is obtained and output.

[0028] Preferably, in step S101, the image to be inspected, which is collected by the infrared camera, is uniformly converted into a grayscale image to simplify the calculation dimension; then, noise filtering is performed in the grayscale image, wherein a 3×3 Gaussian filter is used to suppress random noise, which is expressed as:

[0029] ;

[0030] Where, Used to control the filter strength, represents the neighborhood center, Represents a grayscale image;

[0031] In noise filtering, for the stripe noise caused by equipment vibration, the two-dimensional discrete Fourier transform (DFT) is used to transform the contact network image after grayscale processing. After Fourier transform, the low-frequency part of the image (representing the overall structure and outline of the image) is concentrated in the center of the spectrum, and the high-frequency part (corresponding to the details and noise of the image) is distributed at the edge. The stripe noise caused by equipment vibration will appear as obvious bright spots or peaks in the frequency domain; further, by observing the spectrum diagram after Fourier transform, the coordinates of the bright spots generated by periodic noise are found, and a Notch filter is constructed. The constructed Notch filter is multiplied with the spectrum after Fourier transform to suppress the frequency components corresponding to the noise, which is used to remove the peak part of the spectrum after Fourier transform that represents periodic noise; the inverse Fourier transform is to convert the frequency domain image after filtering back to the spatial domain to obtain an image after removing the stripe noise.

[0032] Preferably, in step S102, the threshold T is calculated using the Otsu algorithm (maximum inter-class variance method), and the image is divided into a dark area and a background based on the threshold T, which is expressed as:

[0033] ;

[0034] Traverse all possible thresholds T and find the between-class variance The maximum threshold T, Represents the background pixel ratio, Indicates the proportion of dark area pixels, 、 Represent the corresponding mean values ​​of background / dark area, Represents the global mean; after image segmentation, the dark area pixels are marked as 1 and the background is marked as 0, and a binary image is obtained. .

[0035] Preferably, in step S103, the Canny operator is used to extract the edge, Gaussian filtering is performed first, and then the gradient amplitude and direction are calculated; non-maximum suppression + double threshold ( 、 ) Filter edges, specifically: Preserve gradient> Strong edges, connecting gradients in − Weak edges of , get edge graph;

[0036] Based on the obtained edge map, contour extraction and screening are performed to obtain the edge contour detection result to verify the wear authenticity of the small wear area, and further determine the wear result of the small wear area based on the verification result.

[0037] Preferably, the components of the gradient in the x-direction and the y-direction are calculated, expressed as:

[0038] ;

[0039] ;

[0040] Where, It represents the gradient component of image I in the horizontal (x) direction, reflecting the grayscale change rate of the image in the horizontal direction; It represents the gradient component of image I in the vertical (y) direction, reflecting the grayscale change rate of the image in the vertical direction; I represents the input grayscale image; and Represent the partial derivatives of image I with respect to x and y, respectively, and are used to calculate the rate of change of the image in the horizontal and vertical directions; 、 is a gradient operator (convolution kernel). The former is a Sobel operator in the horizontal direction, and the latter is a Sobel operator in the vertical direction. By performing convolution operation with the image, the gradient of the image in the corresponding direction is approximately calculated.

[0041] Calculate the gradient magnitude and direction, expressed as: , ;

[0042] Among them, G is the gradient amplitude, which is used to represent the gradient change of the image in the x and y directions, reflecting the drastic degree of grayscale change of the pixel in the image. The larger G is, the more likely the pixel is at the edge; θ is the gradient direction, which indicates the direction of the edge of the image at the pixel point, and is calculated by the inverse tangent function arctan and The gradient direction is obtained by the ratio of , which is used to determine whether the pixel is located at the center of the edge in the non-maximum suppression step, thereby refining the edge.

[0043] Compared with the prior art, the technical advantages of the device for detecting wear of a subway power supply contact network wire according to the embodiment of the present application are:

[0044] First, at the hardware level, the infrared beam emitted by the infrared laser forms the required detection area on the contact network line. The size of the detection area can be set as needed, and then the distance and angle of the infrared laser can be adjusted; the infrared camera is used to collect images of the detection area at different positions. For the non-edge area of ​​the detection area, since the infrared beam of the infrared laser is irradiated on the detection area at an oblique angle, under the irradiation of the infrared beam, if the surface of the detection area of ​​the contact network line has grooves due to wear, the reflection and scattering characteristics of the light in the grooves will change. When the infrared camera is collecting, the groove part will have defective dark areas or grayscale and brightness abnormalities due to the inability of light to return normally or energy differences. These abnormalities can be further identified through image processing algorithms, and the worn area can be marked. When the infrared camera is aimed at the edge of the detection area, the edge contour line image of the detection area can be obtained. Based on the analysis of the edge contour line, the detection result at the edge contour line is obtained.

[0045] In the embodiment of the present application, the detection area is firstly identified for the area with less wear. After the area with less wear is identified, the distance and angle of the infrared laser 4 can be adjusted so that the edge of the new detection area is aligned with the area with less wear detected in the previous step. Then, using edge contour line identification, if the contour line identification of the area with less wear also shows wear, then the area with less wear can be determined to be real wear. On the contrary, if the contour is normal, the area with less wear may be caused by interference (such as stains or accidental light obstruction), thereby improving the detection accuracy. Based on the progressive verification from area to contour line, the accuracy of wear detection can be improved.

[0046] Second, at the software level, the image to be inspected is preprocessed, and the Otsu algorithm is used to divide the preprocessed image into a dark area and a background. After segmentation, a binary image is obtained; based on morphological operations, a morphologically processed binary image is output to highlight continuous dark areas; the morphologically processed binary image is traversed, independent connected domains are marked, and the area, minimum circumscribed rectangle, and grayscale mean of each connected domain are calculated; the grayscale mean is much lower than the background and is identified as a defective dark area, and the coordinates and size are recorded as a small wear area; based on the position of the small wear area, the infrared laser is adjusted so that the edge contour line of the detection area formed by the infrared laser on the contact network line is within the small wear area; and the Canny edge contour detection algorithm is used to verify the authenticity of the wear of the small wear area, and the wear result of the small wear area is determined based on the verification result; based on the wear results of multiple small wear areas, the wear condition of the current section of the contact network line is obtained and output.

[0047] To sum up, the detection device provided in the embodiment of the present application is suitable for the wear detection scenario of subway contact network lines. It can quickly locate defects in the contact network lines, improve accuracy by combining morphology and edge detection, and realize targeted detection through dynamic adjustment of the laser. It effectively solves the technical problem that the detection results of existing detection devices are relatively low in accuracy and it is difficult to meet the needs of accurate detection and analysis of the wear status of the line contact network. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings in the specification, which constitute a part of this application, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0049] In the attached figure:

[0050] Figure 1 This is a schematic structural diagram of a device for detecting wear of a subway power supply contact network line according to an embodiment of the present application;

[0051] Figure 2 A schematic diagram of the internal structure of the detection device provided in an embodiment of the present application;

[0052] Figure 3 A schematic diagram of the detection module provided in an embodiment of the present application fitting within the bottom housing;

[0053] Figure 4 A schematic diagram of the coordination between the contact network wire and the detection module provided in an embodiment of the present application;

[0054] Figure 5 A schematic diagram of the adjustment structure of the infrared laser provided in an embodiment of the present application;

[0055] Figure 6 A schematic diagram of the adjustment structure of the infrared camera provided in an embodiment of the present application;

[0056] Figure 7 A schematic diagram of the detection principle of the contact network line by the detection module provided in an embodiment of the present application;

[0057] Figure 8 A schematic diagram of the implementation flow of the image recognition module provided in an embodiment of the present application.

[0058] The above drawings include the following reference numerals:

[0059] 1. Detection box shell; 11. Box cover; 12. Opening and closing hinge; 13. Limiting ring;

[0060] 2. Contact network line; 21. Detection area;

[0061] 3. Bottom shell;

[0062] 4. Infrared laser; 41. U-shaped frame; 42. First micro servo motor; 43. Limit slider; 44. Guide rail; 45. Electric telescopic rod;

[0063] 5. Infrared camera;

[0064] 6. lateral displacement structure; 61. displacement slider; 62. adjustment screw; 63. guide slide; 64. second micro servo motor; 65. mounting bracket;

[0065] 7. Image processor. DETAILED DESCRIPTION

[0066] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0067] It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meaning as commonly understood by ordinary technicians in the technical field to which this application belongs.

[0068] In the present invention, unless otherwise specified, the directional words used, such as "up, down, top, bottom", usually refer to the directions shown in the drawings, or to the components themselves in the vertical, perpendicular or gravity direction; similarly, for ease of understanding and description, "inside and outside" refer to the inside and outside relative to the outline of each component itself, but the above directional words are not used to limit the present invention.

[0069] like Figure 1-Figure 3 As shown, in one embodiment of the present application, a device for detecting wear of a subway power supply contact network line is provided, which is used to detect surface wear of the contact network line 2;

[0070] Specifically, the wear of the surface where the contact wire 2 and the pantograph are in contact is detected;

[0071] Specifically, at the hardware level, the detection device provided in the embodiment of the present application includes a detection module, which is used to detect the worn surface of the contact wire 2; the detection module includes an infrared laser 4 and an infrared camera 5, wherein, Figure 7 As shown, the infrared laser 4 provided in this embodiment is used to emit an infrared light beam to the surface to be tested of the contact network line 2, and form a detection area 21 on the surface of the contact network line 2, and the infrared camera 5 is used to take a picture of the detection area 21 with infrared light irradiated on the surface to obtain the image to be tested.

[0072] Preferably, the wavelength of infrared laser 4 is in the near-infrared band, which can be 780nm, 808nm, 850nm, 905nm, 980nm, etc.; the infrared camera 5 is equipped with a filter of the corresponding wavelength band. By using infrared laser 4 for imaging detection, the interference of stray light in the visible light region on the imaging is avoided. In principle, the reflection and scattering characteristics of the metal material of the contact network (such as copper alloy) for near-infrared light are significantly different from those of visible light. When near-infrared light illuminates the surface of the contact network, the energy distribution is more consistent with the imaging requirements of wear characteristics (such as grooves and defects), allowing the worn areas to form clear grayscale and brightness differences in the image. At the same time, the infrared camera is equipped with a filter of the corresponding wavelength band to accurately filter out stray light in the visible light region (such as ambient natural light and peripheral lighting of the equipment). If these stray lights enter the imaging system, they are likely to cause image noise and artifacts, interfering with the identification of wear features. The filter combined with the near-infrared laser solution constructs a pure imaging environment of "active light source + narrowband filtering", laying a solid foundation for the algorithm to identify suspected wear areas (such as defective dark areas and areas with missing reflections) at the hardware level, ensuring that the detection results are not affected by fluctuations in the external light environment, and improving the system's adaptability and detection reliability in different time periods and scenarios.

[0073] like Figure 5 As shown, in an optional embodiment, the infrared laser 4 provided in the embodiment of the present application is rotatably arranged at the open end of the U-shaped frame 41 by rotating the support shaft, so that the irradiation direction of the light beam of the infrared laser 4 can be adjusted as needed;

[0074] Preferably, the infrared laser 4 is used to generate a circular area light beam, the central axis of the circular area light beam and the central axis of the contact network line 2 are always in the same plane A, so that after the irradiation angle of the light beam of the infrared laser 4 is adjusted, it can be irradiated on the contact network line 2. There is an angle between the central axis of the circular area light beam and the central axis of the contact network line 2, and the angle is between 30°-60°. During irradiation, a detection area 21 is formed on the contact network line 2.

[0075] Please continue to refer to Figure 5 In this embodiment, a first micro servo motor 42 is provided at the open end of the U-shaped frame 41. The first micro servo motor 42 is used to drive the infrared laser 4 to rotate at a desired angle, thereby adjusting the angle of the infrared beam irradiated by the infrared laser 4. Specifically, the output shaft of the first micro servo motor 42 is connected to a rotating support shaft fixedly connected to the infrared laser 4 via a coupling.

[0076] To precisely control the angle adjustment of the infrared laser 4, a servo control system based on a PID control algorithm is employed to drive the first micro-servo motor 42. The PID control automatically adjusts the motor's drive signal based on the deviation between the set angle and the actual feedback angle. When the deviation is large, the adjustment is increased, while when the deviation is small, fine-tuning is performed, ensuring that the infrared laser 4 accurately reaches the target angle and maintains it stably.

[0077] Furthermore, in order to achieve flexible adjustment of the detection area 21 to further expand or reduce the detection area 21, in this embodiment, the distance between the infrared laser 4 and the contact network line 2 is adjustable. In one adjustable implementation, a limit slider 43 is fixedly mounted on the U-shaped frame 41, and an electric telescopic rod 45 is mounted on the limit slider 43. The telescopic end of the electric telescopic rod 45 is fixedly connected to the interior of the bottom shell 3.

[0078] In addition, a guide rail 44 is fixedly provided inside the bottom shell 3 provided in this embodiment. The guide rail 44 cooperates with the limit slider 43, and the guide rail 44 is used to guide the movement of the limit slider 43; that is, when the telescopic function of the electric telescopic rod 45 is used to adjust the distance between the infrared laser 4 and the contact network line 2, it is used to play a stable guiding role.

[0079] Preferably, the electric telescopic rod 45 is a purchased part, and a product with high screw precision can be selected, such as an electric telescopic rod with ball screw drive, which has higher precision than ordinary screws and can achieve a repeat positioning accuracy of ±0.05mm or even higher.

[0080] Please continue to refer to Figure 4 and Figure 6 In this embodiment, when the infrared camera 5 is used to collect images of the detection area 21, the image collection position of the infrared camera 5 is adjusted by the lateral displacement structure 6 to obtain multiple images, which has obvious advantages in comprehensive detection, accurate analysis, fault diagnosis, etc.;

[0081] Specifically, in one implementation of the lateral displacement structure 6 of the embodiment of the present application, the displacement direction axis of the lateral displacement structure 6 relative to the infrared camera 5 is coplanar with the plane A, and the displacement direction axis is parallel to the central axis of the contact wire 2;

[0082] Among them, the lateral displacement structure 6 can adopt a telescopic rod structure, such as a high-precision electric telescopic rod;

[0083] Specifically, the lateral displacement structure 6 of the embodiment of the present application adopts the following structure: Figure 4 and Figure 6As shown, it includes a displacement slider 61, the infrared camera 5 is mounted on the displacement slider 61, the displacement slider 61 is sleeved on the adjustment screw 62 by a threaded connection, and the rotation of the adjustment screw 62 is driven by a second micro servo motor 64;

[0084] In addition, the lateral displacement structure 6 further includes a guide slide 63, which is slidably disposed on the displacement slider 61 to guide the movement of the displacement slider 61. The guide slide 63 and the adjustment screw 62 are arranged parallel to each other to ensure that the displacement slider 61 can only move but not rotate.

[0085] In addition, it can be understood that the output shaft of the second micro servo motor 64 is driven and connected to one end of the adjusting screw 62, such as by a coupling. The second micro servo motor 64 is the same as the first micro servo motor 42 and can also be precisely controlled based on the PID control algorithm.

[0086] Of course, those skilled in the art may not adopt the lateral displacement structure 6 of the embodiment of the present application, but may reasonably and equidistantly arrange multiple infrared cameras 5 as needed to realize the acquisition of multiple images of the detection area, and the specific details are not limited.

[0087] like Figure 6 As shown, the second micro servo motor 64 is fixed on the mounting bracket 65 ; one end of the guide slide bar 63 is also fixedly mounted on the mounting bracket 65 .

[0088] Please continue to refer to Figure 1 and Figure 2 In the embodiment of the present application, the bottom shell 3 is fixedly arranged on the detection box shell 1. Specifically, the upper end of the bottom shell 3 is an open structure, and the opening at the upper end of the bottom shell 3 is communicated with the bottom opening of the detection box shell 1. The infrared laser 4 and the infrared camera 5 are arranged in the bottom shell 3. The structural arrangement in which the opening at the upper end of the bottom shell 3 is communicated with the bottom opening of the detection box shell 1 facilitates the infrared camera 5 to perform image acquisition and utilizes the infrared laser 4 to irradiate the contact network wire 2 in the detection box shell 1 with an infrared beam;

[0089] In this embodiment, a cover 11 is provided on one side of the detection box shell 1 through an opening and closing hinge 12. The cover 11 is opened to facilitate the insertion of the contact network line 2 into the detection box shell 1, and then the cover 11 is closed.

[0090] Preferably, a limit snap ring 13 is provided on both the detection box shell 1 and the box cover 11, and the limit snap ring 13 adopts a split structure, that is, the limit snap ring 13 is a structure composed of two half rings, and the two half rings are respectively provided on the box cover 11 and the detection box shell 1, wherein the half ring on the detection box shell 1 is a fixed connection, and the half ring on the box cover 11 can be a detachable structure; in specific implementation, when the contact network line 2 is placed in the detection box shell 1, the contact network line 2 is first pressed against a half ring, and when the box cover 11 is closed, the other half ring is used to limit the contact network line 2. Since the contact network line 2 is a circular structure with a notch, the structure of the limit snap ring 13 matches the shape of the contact network line 2, which can prevent the detection box shell 1 from rotating relative to the contact network line 2.

[0091] Furthermore, the detection device provided in the embodiment of the present application further includes an image processor 7, which is used to receive the image to be inspected captured by the infrared camera 5. The image processor 7 has a built-in image recognition module, which is used to identify the wear condition of the detection area 21 based on the image to be inspected;

[0092] like Figure 7 As shown, the embodiment of the present application provides a detection method for the detection device of subway power supply contact network wear based on the above embodiment, in which:

[0093] First, open the box cover 11, place the detection box shell 1 on the contact network line 2, close the box cover 11 and fix it. At this time, the detection box shell 1 is fixed on the contact network line 2;

[0094] Then, the distance of the infrared laser 4 relative to the contact wire 2 and the angle of the infrared laser 4 are adjusted as needed, so that the infrared beam emitted by the infrared laser 4 forms a desired detection area 21 on the contact wire 2. That is, the size of the detection area 21 can be set as needed, and then the distance and angle of the infrared laser 4 can be adjusted.

[0095] Afterwards, the infrared camera 5 is used to collect images of the detection area 21 at different positions. For the non-edge area of ​​the detection area 21, since the infrared beam of the infrared laser 4 is irradiated on the detection area 21 at an oblique angle, under the irradiation of the infrared beam, if the detection area surface of the contact network line 2 has grooves due to wear, the reflection and scattering characteristics of the light in the grooves will change. When the infrared camera 5 collects images, the grooves will have defective dark areas or grayscale and brightness abnormalities due to the inability of light to return normally or energy differences. These abnormalities can be further identified through image processing algorithms, and the worn areas can be marked. When the infrared camera 5 is aimed at the edge of the detection area 21, an edge contour line image of the detection area can be obtained. Based on the analysis of the edge contour line, the detection result at the edge contour line is obtained.

[0096] In particular, in the detection method of the embodiment of the present application, the detection area 21 can be firstly used to identify the area with small wear. After the area with small wear is identified, the distance and angle of the infrared laser 4 can be adjusted so that the edge of the new detection area 21 is aligned with the area with small wear detected in the previous step. By using edge contour line identification, if the contour line identification of the area with small wear also shows wear, then it can be determined that the area with small wear is real wear; on the contrary, if the contour is normal, the area with small wear may be caused by interference (such as stains, accidental light obstruction), thereby improving the detection accuracy; based on the progressive verification from area to contour line, the accuracy of wear detection can be improved.

[0097] In addition, the image recognition module provided in the embodiment of the present application is the software layer of the detection device in the embodiment of the present application;

[0098] Specifically, if Figure 8 As shown, in one implementation of the image recognition module in the embodiment of the present application, the image recognition module is used to implement the following steps:

[0099] S101, pre-processing the image to be inspected, wherein the image to be inspected is one or more images captured by the infrared camera 5;

[0100] S102. Use the Otsu algorithm (maximum inter-class variance method) to divide the pre-processed image into dark areas and background. After segmentation, mark the dark area pixels as 1 and the background pixels as 0, obtaining a binary image. Output the morphologically processed binary image based on morphological operations to highlight continuous dark areas. Traverse the morphologically processed binary image, mark independent connected domains, and calculate the area, minimum bounding rectangle, and grayscale mean of each connected domain. Pixels with an area smaller than a threshold are determined to be noise and filtered. Pixels with a grayscale mean much lower than the background are identified as defective dark areas, and their coordinates and size are recorded as areas of minimal wear.

[0101] Preferably, in the step of dividing the pre-processed image into dark areas and background using the Otsu algorithm, the image is divided into sub-areas of different scales, and the Otsu algorithm is applied to each sub-area, which can effectively deal with the grayscale overlap problem caused by ambient temperature interference and local uneven illumination;

[0102] S103. Based on the location of the small wear area, the infrared laser 4 is adjusted so that the edge contour line of the detection area 21 formed by the infrared laser 4 on the contact wire 2 is within the small wear area; the authenticity of the wear in the small wear area is verified using the Canny edge contour detection algorithm, and the wear result of the small wear area is determined based on the verification result;

[0103] In step S103, the projection area of ​​the infrared laser 4 and the image coordinate system need to be pre-calibrated to ensure the spatial mapping accuracy of the edge contour line of the detection area and the wear area;

[0104] S104 : Based on the wear results of the multiple small wear areas, the wear condition of the current section of the contact network line 2 is obtained and output.

[0105] Furthermore, in step S101, the image to be inspected, which is captured by the infrared camera 5, is uniformly converted into a grayscale image to simplify the calculation dimension; then, noise filtering is performed in the grayscale image, wherein a 3×3 Gaussian filter is used to suppress random noise, which is expressed as:

[0106] ;

[0107] Where, Used to control the filter strength, represents the neighborhood center, Represents a grayscale image;

[0108] In noise filtering, for stripe noise caused by equipment vibration, a two-dimensional discrete Fourier transform (DFT) is used to transform the grayscaled contact network image. After the Fourier transform, the low-frequency part of the image (representing the overall structure and outline of the image) is concentrated in the center of the spectrum, and the high-frequency part (corresponding to the image details and noise) is distributed at the edge. The stripe noise caused by equipment vibration will appear as obvious bright spots or peaks in the frequency domain. Furthermore, by observing the spectrum diagram after the Fourier transform, the coordinates of the bright spots generated by periodic noise are found, and a Notch filter is constructed. The constructed Notch filter is multiplied by the Fourier transformed spectrum to suppress the frequency components corresponding to the noise and remove the peak part representing the periodic noise in the Fourier transformed spectrum. The inverse Fourier transform converts the filtered frequency domain image back to the spatial domain to obtain an image after the stripe noise is removed.

[0109] Furthermore, in step S102, the threshold T is calculated using the Otsu algorithm (maximum inter-class variance method), and the image is divided into dark areas and background based on the threshold T, which is expressed as:

[0110] ;

[0111] Traverse all possible thresholds T and find the between-class variance The maximum threshold T, Represents the background pixel ratio, Indicates the proportion of dark area pixels, 、 Represent the corresponding mean values ​​of background / dark area, Represents the global mean; after image segmentation, the dark area pixels are marked as 1 and the background is marked as 0, and a binary image is obtained. .

[0112] Furthermore, in step S103, the Canny operator is used to extract the edge, Gaussian filtering is performed first, and then the gradient amplitude and direction are calculated, wherein the components of the gradient in the x direction and the y direction are calculated, which are expressed as:

[0113] ;

[0114] ;

[0115] Where, It represents the gradient component of image I in the horizontal (x) direction, reflecting the grayscale change rate of the image in the horizontal direction; It represents the gradient component of image I in the vertical (y) direction, reflecting the grayscale change rate of the image in the vertical direction; I represents the input grayscale image; and Represent the partial derivatives of image I with respect to x and y, respectively, and are used to calculate the rate of change of the image in the horizontal and vertical directions; 、 is a gradient operator (convolution kernel). The former is a Sobel operator in the horizontal direction, and the latter is a Sobel operator in the vertical direction. By performing convolution operation with the image, the gradient of the image in the corresponding direction is approximately calculated.

[0116] The gradient magnitude and direction are further calculated as: , Where G is the gradient amplitude, which is used to represent the gradient change of the image in the x and y directions, reflecting the intensity of the grayscale change of the pixel in the image. The larger G is, the more likely the pixel is at the edge. θ is the gradient direction, which indicates the direction of the edge of the image at the pixel point, and is calculated by the inverse tangent function arctan. and The gradient direction is obtained by the ratio of , which is used to determine whether the pixel is located at the center of the edge in the non-maximum suppression step, thereby refining the edge;

[0117] At the end of using the Canny operator to extract the edge, non-maximum suppression + double threshold ( 、 ) Filter edges, specifically: Preserve gradient> Strong edges, connecting gradients in − Weak edges of , get edge graph;

[0118] Based on the obtained edge map, contour extraction and screening are performed to obtain the edge contour detection result to verify the wear authenticity of the small wear area, and further determine the wear result of the small wear area based on the verification result.

[0119] To sum up, the detection device and detection method provided in the embodiments of the present application can quickly locate defects in the contact network line 2, improve accuracy by combining morphology and edge detection, and realize targeted detection through dynamic adjustment of the laser, which effectively solves the technical problem that the detection results of the existing detection devices are relatively low in accuracy and difficult to meet the needs of accurate detection and analysis of the wear status of the line contact network. It is suitable for wear detection scenarios of subway contact network lines.

[0120] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and exemplary embodiments. They can be applied to a variety of fields suitable for the present invention. Further modifications will be readily apparent to those skilled in the art. Therefore, the present invention is not limited to the specific details and illustrations shown and described herein without departing from the general concept defined by the claims and their equivalents.

Claims

1. A device for detecting wear of a subway power supply contact network line, characterized in that: include: The detection module includes an infrared laser and an infrared camera. The infrared laser is used to emit an infrared beam to the surface to be tested of the contact network line and form a detection area on the surface of the contact network line. The infrared camera is used to take a picture of the detection area to collect the image to be tested. The infrared laser is rotatably arranged at the open end of the U-shaped frame by rotating the support shaft. The infrared laser is used to generate a circular area light beam. The central axis of the circular area light beam and the central axis of the contact network line are always in the same plane A, and the angle between the central axis of the circular area light beam and the central axis of the contact network line is 30°-60°. When using the infrared camera to collect images of the detection area, the image collection position of the infrared camera is adjusted by the lateral displacement structure. The displacement direction axis of the infrared camera of the lateral displacement structure is coplanar with plane A, and the displacement direction axis is parallel to the central axis of the contact network line. An image processor is used to receive the image to be inspected captured by the infrared camera. The image processor has a built-in image recognition module. The image recognition module is used to identify the wear condition of the inspection area based on the image to be inspected. The image recognition module is used to implement the following steps: based on the position of the area with small wear, the infrared laser is adjusted so that the edge contour line of the inspection area formed by the infrared laser on the contact network line is within the area with small wear; and the Canny edge contour detection algorithm is used to verify the authenticity of the wear in the area with small wear, and the wear result of the area with small wear is determined based on the verification result; based on the wear results of multiple areas with small wear, the wear condition of the current section of the contact network line is obtained and output.

2. The device for detecting wear of a subway power supply contact network line according to claim 1, characterized in that: A first micro servo motor is provided on the open end of the U-shaped frame; The output shaft of the first micro servo motor is connected to the rotating support shaft fixedly connected to the infrared laser through a coupling; A limit slider is fixedly installed on the U-shaped frame, an electric telescopic rod is installed on the limit slider, and the telescopic end of the electric telescopic rod is fixedly connected to the inside of the bottom shell; A guide rail is fixedly provided inside the bottom shell, and the guide rail cooperates with the limit slider.

3. The device for detecting wear of a subway power supply contact network line according to claim 2, characterized in that: The lateral displacement structure includes: A displacement slider, the infrared camera is mounted on the displacement slider, the displacement slider is sleeved on the adjustment screw by a threaded connection, and the rotation of the adjustment screw is driven by a second micro servo motor; the second micro servo motor is fixed on the mounting frame; The guide slide rod is slidably arranged on the displacement slider, and the guide slide rod and the adjusting screw rod are arranged parallel to each other; one end of the guide slide rod is also fixedly installed on the mounting frame.

4. The device for detecting wear of a subway power supply contact network line according to claim 2 or 3, characterized in that: The bottom shell is fixedly arranged on the detection box shell, and the upper end of the bottom shell is an open structure; The opening at the upper end of the bottom shell is communicated with the bottom opening of the detection box shell; The infrared laser and the infrared camera are arranged in the bottom shell; A box cover is provided on one side of the detection box shell in an opening and closing manner through an opening and closing hinge; The detection box shell and the box cover are both provided with limit clamps, and the limit clamps adopt a split structure.

5. The device for detecting wear of a subway power supply contact network line according to claim 4, characterized in that: The area with less wear is obtained by the following steps: S101, preprocessing an image to be inspected, wherein the image to be inspected is one or more images captured by an infrared camera; S102. Use the Otsu algorithm to divide the pre-processed image into dark areas and background. After segmentation, the dark area pixels are marked as 1 and the background pixels are marked as 0 to obtain a binary image. Based on morphological operations, the binary image after morphological processing is output to highlight continuous dark areas. The binary image after morphological processing is traversed to mark independent connected domains. The area, minimum circumscribed rectangle, and grayscale mean of each connected domain are calculated: pixels with an area smaller than the threshold are judged as noise and filtered; pixels with a grayscale mean much lower than the background are confirmed as defective dark areas, and the coordinates and size are recorded as small wear areas.

6. The device for detecting wear of a subway power supply contact network line according to claim 5, characterized in that: In step S101, the image to be inspected, which is captured by the infrared camera, is uniformly converted into a grayscale image; then, noise in the grayscale image is filtered; Among them, 3×3 Gaussian filtering is used to suppress random noise, which is expressed as: ； Where, Used to control the filter strength, represents the neighborhood center, Represents a grayscale image.

7. The device for detecting wear of a subway power supply contact network line according to claim 6, characterized in that: In step S102, the threshold T is calculated using the Otsu algorithm, and the image is divided into dark areas and background based on the threshold T, which is expressed as: ； Traverse all possible thresholds T and find the between-class variance The maximum threshold T, Represents the background pixel ratio, Indicates the proportion of dark area pixels, 、 Represent the corresponding mean values ​​of background / dark area, Represents the global mean; after image segmentation, the dark area pixels are marked as 1 and the background is marked as 0, and a binary image is obtained. .

8. The device for detecting wear of a subway power supply contact network line according to claim 7, characterized in that: In step S103, the Canny operator is used to extract edges, Gaussian filtering is performed first, and then the gradient amplitude and direction are calculated; Using non-maximum suppression + double threshold ( 、 ) Filter edges, specifically: Preserve Gradients> Strong edges, connecting gradients in − Weak edges of , get edge graph; Based on the obtained edge map, contour extraction and screening are performed to obtain the edge contour detection result.

9. The device for detecting wear of a subway power supply contact network line according to claim 8, characterized in that: Calculate the components of the gradient in the x and y directions, expressed as: ； ； Where, It represents the gradient component of image I in the horizontal (x) direction, reflecting the grayscale change rate of the image in the horizontal direction; It represents the gradient component of image I in the vertical (y) direction, reflecting the grayscale change rate of the image in the vertical direction; I represents the input grayscale image; and Represent the partial derivatives of image I with respect to x and y, respectively, and are used to calculate the rate of change of the image in the horizontal and vertical directions; 、 The former is the Sobel operator in the horizontal direction, and the latter is the Sobel operator in the vertical direction. By performing convolution operation with the image, the gradient of the image in the corresponding direction is approximately calculated; Calculate the gradient magnitude and direction, expressed as: , ; Among them, G is the gradient amplitude, which is used to represent the gradient change of the image in the x direction and y direction; θ is the gradient direction, which is calculated by the arctan function and The ratio of is used to get the gradient direction.

Images