Wear analysis method of rigid contact network based on binocular vision

Abstract

The present invention provides a wear analysis method for a rigid contact network based on binocular vision. The method comprises: collecting contact line images of the rigid contact network on the top of the train and the top of the tunnel by two cameras; correcting the contact line image pair collected by the two cameras, and obtaining a disparity map of the corrected image pair by using a stereo matching algorithm; converting the disparity map into a depth map according to the principle of binocular vision stereo imaging, extracting the contact line part in the depth map, and performing three-dimensional reconstruction and visualization of the contact line part to obtain the wear characteristics and distribution of the rigid contact network. The present invention utilizes a three-dimensional map of the contact network surface and various types of wear disease curves to better describe the wear of the contact line surface and realize automated and intelligent detection of the rigid contact network.

Description

Technical Field

[0001] The invention relates to the field of urban rail transit system equipment fault diagnosis, and in particular to a binocular vision-based wear analysis method for a rigid contact network. Background Art

[0002] The pantograph-catenary system is a key component of the electrified railway power supply system. It is responsible for the important task of transmitting the traction network power to the electric locomotive for use, and the contact network is a key component of the pantograph-catenary system. During the high-speed operation of urban rail trains, the pantograph and the contact network have always maintained a high-speed friction state. Coupled with the influence of natural environmental erosion, it is easy for the surface of the contact network to suffer from wear and other defects. If these abnormalities or defects are not discovered in time, they may lead to serious consequences and accidents. The contact network is arranged along the train line, and the rigid contact network is installed on the top of the tunnel. The characteristics of the extremely wide length range and high installation height make the detection task of the urban rail transit contact network difficult and the workload is also huge. Therefore, the research and implementation of the automated and intelligent detection scheme of the rigid contact network of urban rail transit is of great significance.

[0003] At present, the relevant research on contact network wear detection in China is basically based on computer vision technology. Different research institutes use different detection schemes, most of which are based on monocular camera schemes. When the contact network is worn, there will generally be a cross section on its bottom surface. The reflectivity of the cross section to light is stronger than that of the rest of the arc part of the contact line. Therefore, in the collected image, the worn edge of the contact network will produce obvious gradient changes. Based on this idea, the monocular camera scheme takes the contact network image from bottom to top, and extracts the contour information of the worn part in one image through image enhancement, image analysis, edge detection and morphological processing methods, and then analyzes the wear of the contact line. In addition to the monocular camera scheme, some studies have adopted binocular vision schemes, which also utilize the characteristics of the worn edge of the bottom of the contact line. Based on binocular ranging technology and triangulated imaging principles, the coordinate information of two points on the bottom edge is obtained to further analyze the bottom wear information.

[0004] At present, the monocular camera solution and the binocular camera solution in the existing technology only use part of the image information. The monocular camera solution uses the wear width and the standard size of the contact line to calculate the wear amount, and the binocular camera solution obtains the coordinate information of two points on the wear edge to calculate the wear amount. Existing solutions all assume that the wear surface of the contact line is smooth and uniform, and cannot obtain other types of wear information. They ignore the characteristic information of the middle wear area and cannot well describe the overall information of the contact line. However, in actual environments, the manifestations of contact line wear are often very diverse.

[0005] Therefore, how to detect the contact network efficiently, intelligently and accurately has become a problem that needs to be solved urgently. Summary of the invention

[0006] The embodiment of the present invention provides a wear analysis method of a rigid contact network based on binocular vision to overcome the problems of the prior art.

[0007] In order to achieve the above object, the present invention adopts the following technical scheme.

[0008] A wear analysis method for rigid contact network based on binocular vision, comprising:

[0009] The contact line images of the rigid contact network on the top of the train and the top of the tunnel are collected by two cameras;

[0010] Correcting the contact line image pair captured by the two cameras, and obtaining a disparity map of the corrected image pair using a stereo matching algorithm;

[0011] The disparity map is converted into a depth map according to the binocular vision stereo imaging principle, the contact line part in the depth map is extracted, the contact line part is three-dimensionally reconstructed and visualized, and the wear characteristics of the rigid contact network are obtained.

[0012] Preferably, before collecting the contact line images of the rigid contact network on the top of the train and the top of the tunnel by two cameras, the method includes:

[0013] A binocular vision system is built according to the binocular vision theory system. The binocular vision system includes two cameras, a device fixing system, a data transmission system and a processing host. The intrinsic parameters, extrinsic parameters and distortion parameters of the camera are calibrated. The camera parameters describe the transformation relationship between the pixel coordinate system and the camera coordinate system. The camera extrinsic parameters describe the transformation relationship between the camera coordinate system and the world coordinate system. The camera distortion parameters are used to correct the image barrel or fisheye effect caused by the camera structural error. The two cameras are installed on the left and right sides of the roof of the rail train and fixed in a forward parallel manner.

[0014] Preferably, the correction of the contact line image pair collected by the two cameras comprises:

[0015] The contact line image pair is corrected using the intrinsic parameters, extrinsic parameters and distortion parameters of the camera. Assume that the extrinsic parameters of the left camera and the right camera are R l 、T l and R r 、T r , where R l 、T l and R r 、T rRepresents the rotation and translation matrices of the transformation between the left camera and the right camera and the same world coordinate system. Given a point P in the world coordinate system, its coordinates in the left and right camera coordinate systems are P l and P r , P l , P r The relationship between it and the world coordinate system is as shown in formula (1).

[0016]

[0017] For the coordinates in the left and right camera coordinate systems, there is a relationship as shown in formula (2).

[0018] P l =R T (P r -T) (2)

[0019] By using the two relations (1) and (2), we can solve the simple relation shown in equation (3).

[0020]

[0021] Solving the above equations, we can obtain the rotation and translation relationships R and T between the left and right cameras. Then, through the Bouguet algorithm in stereo correction, the contact line images captured by the left and right cameras are corrected to be coplanar and aligned with each other, with the goal of minimizing the number of reprojections of each image in the left and right images and maximizing the observation area.

[0022] Preferably, the stereo matching algorithm includes an SGM semi-global matching algorithm.

[0023] Preferably, the converting of the disparity map into a depth map according to the binocular vision stereo imaging principle includes:

[0024] For a point P in the world coordinate system, its coordinates in the left and right camera coordinate systems are P l and P r , the horizontal positions on the left and right camera imaging surfaces are x l and x r , the parallax is the difference between the horizontal coordinates of a point on the left and right camera imaging surfaces, that is, d = x l -x r , using the principle of similar triangles, we get the following formula:

[0025]

[0026] Where T represents the base distance between the left and right cameras, and Z represents the distance between point P and the system baseline.

[0027] After the coordinates of a certain point in the left and right camera coordinate systems and the reference distance between the cameras are known, the distance of point P from the system baseline is calculated according to equation (4), and the depth map is obtained according to the distances of all points from the system baseline.

[0028] Preferably, the extracting of the contact line portion in the depth map, performing three-dimensional reconstruction and visualization on the contact line portion, and obtaining the wear characteristics and distribution of the rigid contact network include:

[0029] Interpolate the missing values ​​in the original disparity map, i.e. the points where the matching fails, and fill them with the median value of the non-zero value in the 3×3 area around the missing value;

[0030] In the depth map, the bottom surface of the contact line is taken as the zero plane, and the depth physical coordinates are converted into pixel coordinates;

[0031] The depth map is filtered and denoised, and local mutations are interpolated and smoothed.

[0032] The mesh function is used in MATLAB to draw a three-dimensional reconstruction of the contact line surface, and the three-dimensional reconstruction is compared with the original contact network image.

[0033] Preferably, the method of obtaining the wear characteristics and distribution of the rigid contact network includes:

[0034] 14 sampling points are set equidistantly in the x-coordinate direction along the contact line section direction, and 14 wear curves in the y-direction are extracted respectively. These 14 wear curves are used to describe different types of contact line diseases respectively. According to the unit ratio relationship between pixels and millimeters in the standard structure diagram and the three-dimensional reconstruction diagram, or a relatively intact section is selected from the result diagram as a reference benchmark without wear, the wear amount at each position on the sampling curve is further obtained, and different diseases are defined:

[0035] Wear ratio

[0036] For a certain cross section, the wear ratio formula is as follows:

[0037] Δw=(h c -h r ) / h c (5)

[0038] Among them, h c Represents the contact line height under perfect conditions, h r Represents the remaining contact line height in the 3D reconstruction. For a wear curve with a length of L, the average wear ratio is

[0039] Eccentric wear

[0040] Eccentric wear means that the contact line is subjected to uneven force on the left and right sides during the contact with the pantograph carbon slide plate, resulting in different degrees of wear on the left and right sides. Using 14 sampled wear curves, for a certain section, when the difference between the wear ratio of the seven curves on the left and the wear ratio of the seven curves on the right is greater than a certain threshold, it means that eccentric wear has occurred. The formula is shown as follows:

[0041]

[0042] Among them, P w Represents the threshold for determining the presence of eccentric wear, Δw i , Δw j分 It indicates the wear value of a certain point on the left and right side of a certain cross section;

[0043] Grooves

[0044] Whether grooves are present is determined by calculating the slope of the wear curve, as shown in the following formula.

[0045]

[0046] Among them, G w represents the threshold for determining the presence of a groove, w(y i+1 ), w(y i ) represents the wear value of two points in the y direction.

[0047] It can be seen from the technical solutions provided by the above-mentioned embodiments of the present invention that the embodiments of the present invention utilize the three-dimensional graph of the contact network surface and various types of wear disease curve graphs to better describe the wear conditions of the contact line surface and realize automated and intelligent detection of the rigid contact network.

[0048] Additional aspects and advantages of the present invention will be given in part in the following description, which will become obvious from the following description, or may be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0050] Figure 1 A processing flow chart of a method for analyzing wear of a rigid contact network based on binocular vision provided by an embodiment of the present invention;

[0051] Figure 2 A schematic diagram of a pair of partial chessboard images taken by left and right cameras provided in an embodiment of the present invention;

[0052] Figure 3 A schematic diagram of a calibration reprojection error provided by an embodiment of the present invention;

[0053] Figure 4 A schematic diagram of a pair of overhead contact network images taken by left and right cameras provided in an embodiment of the present invention;

[0054] Figure 5 A schematic diagram of a corrected contact line image pair provided by an embodiment of the present invention;

[0055] Figure 6 A disparity map obtained by stereo matching provided by an embodiment of the present invention;

[0056] Figure 7 A schematic diagram of a stereoscopic imaging principle provided by an embodiment of the present invention;

[0057] Figure 8 A schematic diagram of a reconstruction result after processing (left) and a comparison with the original image (right) provided in an embodiment of the present invention;

[0058] Fig. 9 A schematic diagram of a detection sampling curve provided by an embodiment of the present invention;

[0059] Fig.10 A schematic diagram of a partial curve wear ratio provided by an embodiment of the present invention;

[0060] Fig.11 A schematic diagram of contact line eccentric wear provided by an embodiment of the present invention;

[0061] Fig.12 A schematic diagram of a contact line groove provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0063] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The term "and / or" used herein includes any unit and all combinations of one or more associated listed items.

[0064] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as herein.

[0065] To facilitate understanding of the embodiments of the present invention, several specific embodiments will be further explained below with reference to the accompanying drawings, and each embodiment does not constitute a limitation on the embodiments of the present invention.

[0066] Embodiment 1

[0067] In view of the problems of the prior art, the embodiment of the present invention is based on binocular stereo vision theory, and on the basis of binocular ranging, after system calibration, image correction, stereo matching, three-dimensional reconstruction and other steps, after adjustment, optimization and conversion, the complete three-dimensional information of the bottom surface of the contact line is obtained, and the three-dimensional structure diagram of the contact network surface is obtained. At the same time, a mathematical definition method of wear disease based on sampling wear curve is proposed for three-dimensional data, and the wear state of the contact network is quantitatively analyzed. From three-dimensional information display to mathematical quantitative analysis, more intuitive results are provided to the subway maintenance department, providing technical support for subway operation and maintenance.

[0068] The processing flow of a wear analysis method for a rigid contact network based on binocular vision proposed in an embodiment of the present invention is as follows: Figure 1 The process comprises the following steps:

[0069] Step S10: System construction and parameter calibration.

[0070] According to the binocular vision theory system, a binocular vision system is built. The system mainly includes two cameras, a device fixing system, a data transmission system and a processing host. Before use, the camera parameters need to be calibrated, mainly including intrinsic parameters, extrinsic parameters and distortion parameters. There are four commonly used coordinate systems in the camera linear model, including: pixel coordinate system, image coordinate system, camera coordinate system and world coordinate system. The camera intrinsic parameters describe the transformation relationship between the pixel coordinate system and the camera coordinate system. The camera extrinsic parameters describe the transformation relationship between the camera coordinate system and the world coordinate system. The distortion parameters are used to correct the "barrel" or "fisheye" effect of the image caused by the camera structure error. Using the calibrated parameters, the collected images can be corrected to obtain more accurate results.

[0071] The above two cameras can be installed on the left and right sides of the roof of the rail train.

[0072] Step S20: Correct the contact line image pair collected by the left and right cameras, and use a stereo matching algorithm to obtain a disparity map of the corrected image pair.

[0073] The contact line image pairs collected by the left and right cameras are corrected using the system parameters obtained by system calibration. The corrected image pairs are distortion-free and aligned with each other. The disparity map of the corrected image pairs is obtained using a stereo matching algorithm. According to the principle of triangulated imaging, the depth map of the scene can be obtained using the disparity map, that is, the distance between each point on the contact line surface and the system is obtained. Further, the contact line surface can be reconstructed and visualized in three dimensions through the depth map, and the wear information can be analyzed.

[0074] Step S30: 3D reconstruction and visualization.

[0075] Using the disparity map of the image pair obtained in the previous step, the disparity map is converted into a depth map according to the principle of binocular vision stereo imaging, so that the wear information of the contact line surface can be obtained. After a series of optimizations are performed on the depth map and the contact line part is extracted, the contact line surface can be reconstructed and visualized in three dimensions using visualization tools, and the wear characteristics and distribution of the contact network can be intuitively observed.

[0076] Step S40: Wear analysis.

[0077] In order to provide more intuitive quantitative wear data information, a mathematical definition method for wear disease based on sampling wear curves was proposed, the wear state of the contact network was classified and defined and the standards were given, and an example analysis was carried out on the photographed contact line images.

[0078] Embodiment 2:

[0079] Step S10: System construction and parameter calibration

[0080] This system uses two identical industrial-grade cameras, model Basler acA1300-60gm, with a resolution of 1280×1024 and a lens with a fixed focus of 25mm. The camera is fixed with a camera bracket, and the position and angle parameters are adjusted. The optical center distance between the left and right cameras is 80mm, and they are fixed in a forward parallel manner. In a laboratory environment, the device is installed on the simulated car roof and the contact network pictures are taken upward. After the system is built, the system needs to be calibrated. The calibration parameters include internal parameters, external parameters and distortion parameters.

[0081] The camera intrinsic parameters describe the transformation relationship between the pixel coordinate system and the camera coordinate system. The intrinsic parameter matrix is f x and f y is the pixel focal length of the camera, which is actually the physical focal length f of the camera and the size s of each unit on the imager x (representing the number of pixels per millimeter, in units of pixel / mm), through c x and c y Represents the possible offset of the optical axis; the camera extrinsic parameters describe the transformation relationship between the camera coordinate system and the world coordinate system. The world coordinate system here usually refers to the world coordinate system where the calibration object is located, which can be described by a rotation and translation matrix; the distortion parameters are used to correct the "barrel" or "fisheye" effect of the image caused by the camera structure error, including radial distortion and tangential distortion, and are described by a 5×1 distortion vector.

[0082] The method used to calibrate the system is the "Zhang Zhengyou calibration method". During the calibration process, only a printed black and white chessboard target is required. Figure 2 A schematic diagram of a pair of partial chessboard images taken by left and right cameras according to an embodiment of the present invention is shown in FIG. Figure 2 As shown in the figure, it is easy to operate and has high accuracy. Therefore, it has become one of the most practical and widely used camera calibration methods in computer vision research. The "Zhang Zhengyou calibration method" is a two-step calibration method. The first step is to make a preliminary estimate of the camera parameters, and the second step is to iteratively refine the parameters using the optimization function.

[0083] According to the above content, use the left and right cameras of the system to simultaneously shoot the printed chessboard, then change the chessboard target position and shoot again. In the actual calibration process, in order to obtain more accurate results, usually at least 10 pictures should be taken. The more image pairs, the smaller the possible error. During the experiment, a total of 20 pairs of chessboard images were taken. Some of the chessboard image pairs taken are as follows: Figure 2As shown in Table 1. Using the MATLAB stereo calibration tool, the camera parameters are obtained as shown in Table 1. The external parameters of the chessboard target are different each time it moves. There are a total of 20 external parameters, which will not be shown one by one here.

[0084] Table 1 Camera parameter table

[0085]

[0086] Although iterative optimization is performed during the calibration process, there is still a certain difference between the pixel coordinates calculated by reprojecting the 3D coordinates of the corner points and the pixel coordinates of the initially extracted corner points, namely the reprojection error value. This parameter reflects the accuracy of the calibration result to a certain extent. The reprojection error of all corner points in each image is as follows: Figure 3 As shown, the final overall reprojection error is 0.32 pixels, indicating that the calibration accuracy is high.

[0087] Step S20: Image correction and stereo matching

[0088] After the system is calibrated, the device is installed on the roof of a simulated train in the laboratory, and the rigid contact line between the top of the train and the top of the tunnel is photographed upward. The contact line image obtained is Figure 4 As shown. The system parameters calculated in the previous step are used to correct the image first to eliminate various distortions. In addition, in the field of binocular stereo vision, in order to better perform stereo matching, it is necessary to ensure that the left and right images are coplanar and aligned. For this purpose, the image pair needs to be stereo corrected. Assume that the external parameters of the two left and right cameras in the system are R l 、T l and R r 、T r , where R l 、T l and R r 、T r Represent the rotation and translation matrices of the transformation between the left camera and the right camera and the same world coordinate system. These parameters can be obtained from the external parameters of the system obtained by previous calibration. Given a point P in the world coordinate system, its coordinates in the left and right camera coordinate systems are P l and P r , which has the relationship with the world coordinate system as shown in formula (1).

[0089]

[0090] For the coordinates in the left and right camera coordinate systems, there is a relationship as shown in formula (2).

[0091] P l =R T (P r -T) (2)

[0092] By using the two relations (1) and (2), we can solve the simple relation shown in equation (3).

[0093]

[0094] Solving the above equations can obtain the rotation and translation relationship R and T between the two left and right cameras. Then, through the Bouguet algorithm in stereo correction, the number of reprojections of each left and right image is minimized (which is also the minimization of reprojection distortion), while maximizing the observation area. Finally, the left and right images are corrected to be coplanar and aligned. Figure 5 A schematic diagram of a corrected contact line image pair provided by an embodiment of the present invention.

[0095] After obtaining the corrected image pair, the image can be stereo matched to obtain the disparity map. The disparity is the difference between the horizontal coordinates of a point on the left and right camera imaging planes. Stereo matching is a key step in binocular stereo vision. The image disparity map is obtained through stereo matching, and then the depth map can be obtained, that is, the distance between each point on the contact line surface and the system is obtained. The wear of the contact line surface can be measured through the distance information. The stereo matching method used here is the semi-global matching algorithm (Semi-Global Matching, SGM). The main idea of ​​the SGM algorithm is to use mutual information (MI) to perform single-pixel matching, and to use multiple one-dimensional constraints to approximate global and two-dimensional smooth constraints, that is, to achieve global optimization through local optimization of multi-directional scan lines. The SGM algorithm is divided into several key steps. The first is to calculate the matching cost at the pixel level, followed by cost aggregation based on the global energy function, and finally the calculation of the disparity map and the refinement of the results. On the basis of the SGM algorithm matching, a weighted least squares filter is used to optimize the results, and the final disparity map obtained by stereo matching is as follows Figure 6 As shown. Figure 6 It can be found that the disparity map obtained by matching the contact line part is good, and the wear characteristics of the contact line surface have been well extracted. However, for the bus and ceiling areas, the disparity map results are relatively confusing. This is mainly because the ceiling area is not in focus and the bus part has severe reflections.

[0096] Step S30: 3D reconstruction visualization

[0097] A schematic diagram of a stereoscopic imaging principle provided by an embodiment of the present invention is as follows Figure 7 As shown, for a point P in the world coordinate system, its coordinates in the left and right camera coordinate systems are P l and P r The corresponding horizontal coordinates are x land x r , which represent the horizontal positions on the left and right camera imaging surfaces respectively. Stereo imaging is based on the principle of triangulation. Here, the concept of parallax is introduced. Parallax is the difference between the horizontal coordinates of a point on the left and right camera imaging surfaces, that is, d = x l -x r ,according to Figure 7 , using the principle of similar triangles, it is easy to derive the following formula.

[0098]

[0099] Where T represents the base distance between the left and right cameras, and Z represents the distance between point P and the system baseline.

[0100] After the coordinates of a certain point in the left and right camera coordinate systems and the reference distance between the cameras are known, the distance of point P from the system baseline is calculated according to equation (4), and the depth map is obtained according to the distances of all points from the system baseline.

[0101] according to Figure 5 The disparity map is extracted, and the contact line part with good matching effect is extracted from the disparity map, and the following steps are performed:

[0102] Interpolate the missing values ​​in the original disparity map, i.e. the points where the matching fails, and fill them with the median value of the non-zero value in the 3×3 area around the missing value;

[0103] The depth map is calculated using formula 4, with the bottom surface of the contact line as the zero plane, and the depth physical coordinates are converted into pixel coordinates for the sake of uniformity;

[0104] Filter and denoise the depth map, perform interpolation and smoothing on some local mutations, and make some adjustments to other details.

[0105] The mesh function is used in MATLAB to draw a three-dimensional reconstruction of the contact line surface. Figure 8 The displayed 3D reconstruction is compared with the original contact network image. Figure 8 A schematic diagram of a reconstruction result after processing (left) and a comparison with the original image (right) provided in an embodiment of the present invention.

[0106] Step S40: Wear analysis

[0107] After obtaining a more intuitive 3D reconstruction of the contact line surface and clearly observing the wear characteristics of the surface, some mathematical representation methods of the spatial characteristics of the disease are designed for the 3D structure. The computational complexity of the entire 3D surface is large and it is not easy to define mathematically. In order to better describe the characteristics of the wear surface, 14 sampling points are set equidistantly in the x-coordinate direction, and 14 wear curves in the y-direction are extracted, such as Fig. 9 As shown. These 14 wear curves are used to describe different types of contact line diseases. According to the standard structure diagram and the unit ratio of pixels and millimeters in the three-dimensional reconstruction diagram, or selecting a relatively intact section from the result diagram, it can be used as a reference for no wear, and the wear amount at each position on the sampling curve can be further obtained, so as to define different diseases.

[0108] Wear ratio

[0109] For a certain cross section, the wear ratio formula is as follows:

[0110] Δw=(h c -h r ) / h c (5)

[0111] Among them, h c Represents the contact line height under perfect conditions, h r Represents the remaining contact line height in the 3D reconstruction. For a wear curve with a length of L, its average wear ratio is

[0112] Eccentric wear

[0113] Eccentric wear means that the contact line is subjected to uneven force on the left and right sides for a long time during the contact process with the pantograph carbon plate, resulting in different degrees of wear on the left and right sides, which will accelerate the wear of the pantograph and affect the normal current collection. In order to define eccentric wear, 14 sampled wear curves are used. For a certain section, when the difference between the wear ratio of the seven curves on the left and the wear ratio of the seven curves on the right is greater than a certain threshold, it means that eccentric wear has occurred. The formula is shown as follows:

[0114]

[0115] Among them, P w Indicates the threshold for judging the presence of eccentric wear.

[0116] Grooves

[0117] When a certain position of a wear curve suddenly drops or rises like a cliff, it means that there is a groove disease on the contact line surface. Grooves on the contact line surface are not common. They mainly exist when the contact line itself is not strong enough or the impact is too large, causing cracks or even falling pieces of the contact line, thus forming grooves. Contact line grooves can cause problems such as bow pulling, affecting the normal contact and current collection between the pantograph carbon slide and the contact network. For grooves, you can simply calculate the slope of the wear curve to judge, as shown in the following formula.

[0118]

[0119] Among them, G w Indicates the threshold for determining whether a groove exists.

[0120] For the three diseases defined above, the reconstruction results are calculated separately and the results are shown below.

[0121] Wear ratio

[0122] A schematic diagram of a wear ratio of a partial curve provided by an embodiment of the present invention is as follows: Fig.10 As shown. Fig.10 It can be seen that the two sampling curves are more severely worn at the ends. There is a groove in the interval of pixel coordinates [100,200] for the fourth sampling curve. The wear is most serious at the groove, and a peak value appears. There is a negative wear ratio in the interval of pixel coordinates [600,700]. It can be considered that the height is too high due to local matching errors.

[0123] Eccentric wear

[0124] An embodiment of the present invention provides a diagram of eccentric wear within a range of length L as shown in FIG. Fig.11 As shown. Fig.11 It can be seen that the wear on the left side of the contact line is higher than that on the right side. In most intervals, the average wear ratio on the left side is larger than that on the right side, and it is particularly obvious in the first half of the interval, which can be observed intuitively from the original picture. The eccentric wear threshold set here is 0.5mm.

[0125] Grooves

[0126] Select one of the wear curves with more severe wear and calculate the derivative of each point. Fig.12 A schematic diagram of a contact line groove provided by an embodiment of the present invention, comprising Fig.12 It can be seen that within the pixel interval [100,200], the wear derivative of the wear curve changes extremely drastically, with a minimum and maximum value, and the change interval is small, indicating that there is an obvious groove disease. When the groove judgment threshold is set to ±1, the groove phenomenon in other areas is relatively less severe, and the wear curve changes more smoothly.

[0127] To sum up, the embodiments of the present invention mainly study the three-dimensional reconstruction and wear measurement technology of the rigid contact network surface of urban rail transit based on binocular vision theory. By using the three-dimensional map of the contact network surface and the curve diagrams of various types of wear diseases, the wear condition of the contact line surface can be better described, and the automatic and intelligent detection of the rigid contact network can be realized. It can provide decision-making support for the maintenance and inspection work of the subway maintenance department and provide technical support for the construction of the future comprehensive detection system of urban rail transit.

[0128] Those skilled in the art can understand that the accompanying drawings are only schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0129] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present invention or certain parts of the embodiments.

[0130] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0131] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A wear analysis method for rigid contact network based on binocular vision, characterized in that: include: The contact line images of the rigid contact network on the top of the train and the top of the tunnel are collected by two cameras; Correcting the contact line image pair captured by the two cameras, and obtaining a disparity map of the corrected image pair using a stereo matching algorithm; The disparity map is converted into a depth map according to the binocular vision stereo imaging principle, the contact line part in the depth map is extracted, and the contact line part is three-dimensionally reconstructed and visualized to obtain the wear characteristics of the rigid contact network; The method of obtaining the wear characteristics of the rigid contact network includes: 14 sampling points are set equidistantly in the x-coordinate direction along the contact line section direction, and 14 wear curves in the y-direction are extracted respectively. These 14 wear curves are used to describe different types of contact line diseases respectively. According to the unit ratio relationship between pixels and millimeters in the standard structure diagram and the three-dimensional reconstruction diagram, or a relatively intact section is selected from the result diagram as a reference benchmark without wear, the wear amount at each position on the sampling curve is further obtained, and different diseases are defined: Wear ratio For a certain cross section, the wear ratio formula is as follows: Δw=(h c -h r ) / hour c (5) Among them, h c Represents the contact line height under perfect conditions, h r Represents the remaining contact line height in the 3D reconstruction. For a wear curve with a length of L, the calculus expression of the average wear ratio is: Eccentric wear Eccentric wear means that the contact line is subjected to uneven force on the left and right sides during the contact with the pantograph carbon slide plate, resulting in different degrees of wear on the left and right sides. Using 14 sampled wear curves, for a certain section, when the difference between the wear ratio of the seven curves on the left and the wear ratio of the seven curves on the right is greater than a certain threshold, it means that eccentric wear has occurred. The formula is shown as follows: Among them, p w Represents the threshold for determining the presence of eccentric wear, Δw i , Δw j Respectively represent the wear value of a point on the left and right side of a certain cross section; Grooves The presence of grooves can be determined by calculating the slope of the wear curve, as shown in the following formula: Among them, G w represents the threshold for determining the presence of a groove, w(y i+1 ), w(y i ) represents the wear value of two points in the y direction.

2. The method according to claim 1, characterized in that Before collecting the contact line images of the rigid contact network on the top of the train and the top of the tunnel by two cameras, the following steps are included: A binocular vision system is built according to the binocular vision theory system. The binocular vision system includes two cameras, a device fixing system, a data transmission system and a processing host. The intrinsic parameters, extrinsic parameters and distortion parameters of the camera are calibrated. The intrinsic parameters of the camera describe the transformation relationship between the pixel coordinate system and the camera coordinate system. The extrinsic parameters of the camera describe the transformation relationship between the camera coordinate system and the world coordinate system. The distortion parameters of the camera are used to correct the image barrel or fisheye effect caused by the camera structure error. The two cameras are installed on the left and right sides of the roof of the rail train and are fixed in a forward parallel manner.

3. The method according to claim 2, characterized in that The correcting of the contact line image pair collected by the two cameras comprises: The contact line image pair is corrected using the intrinsic parameters, extrinsic parameters and distortion parameters of the camera. Assume that the extrinsic parameters of the left camera and the right camera are R l 、T l and R r 、T r , where R l 、T l and R r 、T r Represents the rotation and translation matrices of the transformation between the left camera and the right camera and the same world coordinate system. Given a point P in the world coordinate system, its coordinates in the left and right camera coordinate systems are P l and P r , P l , P r The relationship with the world coordinate system is as shown in formula (1); For the coordinates in the left and right camera coordinate systems, there is a relationship as shown in formula (2); P l =R T (P r -T) (2) By using the two relations (1) and (2), we can solve the simple relation shown in equation (3); Solving the above equations, we can obtain the rotation and translation relationships R and T between the left and right cameras. Then, through the Bouguet algorithm in stereo correction, the contact line images captured by the left and right cameras are corrected to be coplanar and aligned with each other, with the goal of minimizing the number of reprojections of each image in the left and right images and maximizing the observation area.

4. The method according to claim 3, characterized in that The stereo matching algorithm includes an SGM semi-global matching algorithm.

5. The method according to claim 1 or 2 or 3 or 4, characterized in that: The converting of the disparity map into a depth map according to the binocular vision stereo imaging principle includes: For a point P in the world coordinate system, its coordinates in the left and right camera coordinate systems are P l and P r , the horizontal positions on the left and right camera imaging surfaces are x l and x r , the parallax is the difference between the horizontal coordinates of a point on the left and right camera imaging surfaces, that is, d = x l -x r , using the principle of similar triangles, we get the following formula: Among them, T represents the base distance between the left and right cameras, and Z represents the distance between point P and the system baseline; After the coordinates of a certain point in the left and right camera coordinate systems and the reference distance between the cameras are known, the distance of point P from the system baseline is calculated according to equation (4), and the depth map is obtained according to the distances of all points from the system baseline.

6. The method according to claim 5, characterized in that The step of extracting the contact line portion in the depth map, performing three-dimensional reconstruction and visualization on the contact line portion, and obtaining the wear characteristics of the rigid contact network includes: Interpolate the missing values ​​in the original disparity map, i.e. the points where the matching fails, and fill them with the median value of the non-zero value in the 3×3 area around the missing value; In the depth map, the bottom surface of the contact line is taken as the zero plane, and the depth physical coordinates are converted into pixel coordinates; The depth map is filtered and denoised, and local mutations are interpolated and smoothed. The mesh function is used in MATLAB to draw a three-dimensional reconstruction of the contact line surface, and the three-dimensional reconstruction is compared with the original contact network image.

Images