A method for calculating the wear of a pantograph slide plate, a storage medium and a computer device

Abstract

The present invention relates to the field of pantograph slide wear calculation, and specifically to a pantograph slide wear calculation method, storage medium and computer equipment. The method utilizes a slide extraction algorithm and a slide reference line extraction algorithm to achieve accurate extraction of the pantograph slide, and utilizes a deep learning registration algorithm and a 3D precision registration algorithm to achieve a high-precision registration effect. The pantograph slide wear calculation method provided by the present invention has a simple algorithm flow involved, and the detection algorithm based on three-dimensional data can eliminate interference from ambient light and slide edge, thereby improving the wear calculation accuracy and stability.

Description

Technical Field

[0001] The invention relates to the field of pantograph slide plate wear calculation, and in particular to a pantograph slide plate wear calculation method, storage medium and computer equipment. Background Art

[0002] The pantograph slide of the train is always in high-speed motion. Any slight damage to the pantograph-network system, such as excessive wear, step wear, notches, grooves, etc. of the pantograph slide, will not only affect the normal power supply of the train, but the resulting arc discharge will further aggravate the wear of the pantograph slide and the contact network, and will also generate radio interference, leading to pantograph-network failure, causing serious losses and impacts on the transportation and production of the rail transit system. Therefore, online detection of pantograph slide wear and mastering the wear law of the slide are of great significance in ensuring the normal operation of locomotives and power grid safety.

[0003] In the prior art, for the wear detection of train pantograph slides, the current common solution is to use a 2D camera to capture images. This technology usually captures the two-dimensional image of the pantograph slide through an area array CCD, extracts the edge of the carbon slide through edge extraction, morphological image processing methods, etc., and realizes the carbon slide wear and residual wear detection through camera calibration. This method is easily affected by ambient light and the edge of the slide, has poor stability, and cannot achieve stable high-precision measurement. Summary of the invention

[0004] In order to solve the problems existing in the above background technology, the present invention proposes a method for calculating the wear of a pantograph slide plate, the method comprising:

[0005] Acquire pantograph three-dimensional data, convert the pantograph three-dimensional data into a pantograph depth map, and extract a reference line of a slide plate to be tested from the pantograph depth map;

[0006] Acquire three-dimensional data of a standard skateboard, convert the three-dimensional data of the standard skateboard into a standard skateboard depth map, and extract a standard skateboard reference line from the standard skateboard depth map;

[0007] The reference line of the slide to be tested is three-dimensionally registered with the reference line of the standard slide to obtain the difference between the reference line of the slide to be tested and the reference line of the standard slide, and the pantograph wear is calculated based on the difference.

[0008] Furthermore, the pantograph three-dimensional data is collected, the pantograph three-dimensional data is converted into a pantograph depth map, and the reference line of the slide plate to be tested is extracted from the pantograph depth map as follows:

[0009] Acquire pantograph three-dimensional data, and convert the pantograph three-dimensional data into a pantograph depth map;

[0010] extracting a slide plate depth map from the pantograph depth map according to a slide plate extraction algorithm;

[0011] The reference line of the skateboard to be tested is extracted from the skateboard depth map according to the skateboard reference line extraction algorithm.

[0012] Further, the step of extracting the slide depth map from the pantograph depth map according to the slide extraction algorithm is specifically as follows:

[0013] Preprocessing the pantograph depth map, eliminating non-slide plate areas, and obtaining a first pantograph depth map;

[0014] Searching for a connected domain on the first pantograph depth map to obtain a connected domain including two slide plates and the middle connecting axes corresponding to the two slide plates;

[0015] Removing the middle connecting axis of the connected domain to obtain a second pantograph depth map including the two slide plates;

[0016] The second pantograph depth map slide is classified to obtain a third pantograph depth map including the slide depth map.

[0017] Further, the step of extracting the reference line of the skateboard to be tested from the skateboard depth map according to the skateboard reference line extraction algorithm is specifically as follows:

[0018] Extracting upper and lower edges of the slide depth map, performing polynomial fitting on the upper and lower edges, and obtaining upper and lower edge curves of the slide in the slide depth map;

[0019] Obtaining upper and lower edge positions of the skateboard in the skateboard depth map through upper and lower edge curves of the skateboard in the skateboard depth map, and then obtaining the center coordinates of the skateboard;

[0020] Performing a straight line fitting on the center coordinates of the skateboard to obtain a straight line expression of the skateboard reference line;

[0021] The coordinates of the skateboard reference line in the skateboard depth map are obtained through the straight line expression, and the three-dimensional data of the skateboard reference line is obtained through the mapping relationship, thereby obtaining the skateboard reference line to be tested.

[0022] Furthermore, the acquisition of the standard skateboard three-dimensional data, the conversion of the standard skateboard three-dimensional data into a standard skateboard depth map, and the extraction of the standard skateboard reference line from the standard skateboard depth map are specifically as follows:

[0023] Repeat the step of extracting the reference line of the skateboard to be tested based on the pantograph depth map in claim 4, and extract the reference line of the standard skateboard based on the standard skateboard depth map.

[0024] Furthermore, the reference line of the skateboard to be tested is three-dimensionally registered with the reference line of the standard skateboard to obtain the difference between the reference line of the skateboard to be tested and the reference line of the standard skateboard, and the pantograph wear is calculated according to the difference as follows:

[0025] Performing deep learning registration on the reference line of the skateboard to be tested and the reference line of the standard skateboard to obtain a first calibrated reference line of the skateboard to be tested;

[0026] Performing 3D precise registration on the first calibration test skateboard reference line to obtain a second calibration test skateboard reference line;

[0027] The pantograph wear is calculated based on the difference between the reference line of the second calibration slide to be tested and the reference line of the standard slide.

[0028] Furthermore, the deep learning alignment of the reference line of the skateboard to be tested and the reference line of the standard skateboard is performed to obtain the first calibrated reference line of the skateboard to be tested, which is specifically:

[0029] Obtaining standard skateboard reference line point cloud data, rotating and translating the standard skateboard reference line point cloud data, adding noise to the connected domain of the middle connecting axes corresponding to the two skateboards, and obtaining a deep learning registration training set;

[0030] According to the deep learning registration training set, the network parameters are updated by using the MSE loss function to perform reverse gradient calculation to obtain a deep learning registration training network;

[0031] Acquire the point cloud data of the reference line of the skateboard to be tested, input the point cloud data of the reference line of the standard skateboard and the point cloud data of the reference line of the skateboard to be tested into the deep learning registration training network, and obtain the first rotation amount and the first translation amount of the reference line of the skateboard to be tested relative to the standard skateboard reference line;

[0032] The first rotation amount and the first translation amount are calculated to obtain the point cloud data of the reference line of the first calibrated skateboard to be tested, and the first calibrated skateboard reference line to be tested is obtained according to the mapping relationship.

[0033] Furthermore, the 3D precise registration of the first calibration skateboard reference line to be tested is performed to obtain the second calibration skateboard reference line to be tested, which is specifically:

[0034] L1: performing filtering and noise reduction processing on the first calibration slide reference line to be tested to obtain the first calibration slide reference line to be tested after noise reduction;

[0035] L2: Calculating the distance between the first calibration test skateboard reference line after noise reduction and the corresponding point of the standard skateboard reference line;

[0036] Setting a threshold, identifying the portion of the distance greater than the threshold as a wear interval, and removing the wear intervals of the first calibration test slide reference line and the standard slide reference line after noise reduction;

[0037] L3: According to formula (1), the first calibration test slide reference line and the standard slide reference line after the wear interval is removed are calculated to obtain the distance from the first calibration test slide reference line after the wear interval is removed to the standard slide reference line after the wear interval is removed.

[0038]

[0039] Among them, x1, y1, z1 represent the standard reference line data, x2, y2, z2 represent the reference line data after deep learning registration, and r 11 -r 33 represents the amount of rotation, t x , t y , t z They represent the translation vectors respectively;

[0040] L4: Iterate L2 and L3. When the distance from the first calibrated skateboard reference line after the wear interval is removed to the standard skateboard reference line after the wear interval is removed is less than a set threshold, stop iteration to obtain the second calibrated skateboard reference line.

[0041] A storage medium stores a computer program, which implements any of the above methods when executed by a processor.

[0042] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements any one of the above methods when executing the program.

[0043] Beneficial effects: The present invention provides a method for calculating the wear of a pantograph slide. The slide extraction algorithm and the slide reference line extraction algorithm are used to accurately extract the pantograph slide. The deep learning registration algorithm and the 3D precision registration algorithm are used to achieve high-precision registration effects. The method for calculating the wear of a pantograph slide provided by the present invention involves a simple algorithm flow. At the same time, the detection algorithm based on three-dimensional data can eliminate interference from ambient light and the edge of the slide, thereby improving the accuracy and stability of wear calculation. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A schematic diagram of a flow chart of a method for calculating the wear of a pantograph slide plate provided in an embodiment of the present invention;

[0045] Figure 2 A three-dimensional diagram of a pantograph provided in an embodiment of the present invention;

[0046] Figure 3 A pantograph point cloud depth map provided by an embodiment of the present invention;

[0047] Figure 4 A point cloud diagram of a skateboard reference line provided by an embodiment of the present invention;

[0048] Figure 5 A curve diagram of the upper surface of a skateboard provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] The present application provides a method for calculating the wear of a pantograph slide, wherein the computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data of the classification method. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for calculating the wear of a pantograph slide is implemented.

[0051] The embodiments of the present application provide a method for calculating the wear of a pantograph slide, a computer device and a storage medium, aiming to solve the problem in the prior art of using a 2D camera to collect images, collecting two-dimensional images of the pantograph slide through an area array CCD, extracting the edge of the carbon slide through edge extraction and morphological image processing methods, and realizing carbon slide wear and residual wear detection through camera calibration, which is easily affected by ambient light and the edge of the slide, has poor stability, and cannot achieve stable high-precision measurement.

[0052] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems will be described in detail below through embodiments and in combination with the accompanying drawings. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. It should be noted that a method for calculating the wear of a pantograph slide provided in an embodiment of the present application, wherein its executor can also be a classification device, wherein the device can be implemented as part or all of the classification method through software, hardware, or a combination of software and hardware.

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0054] In one embodiment, Figure 1 and Figure 2 As shown, when a train passes by, the three-dimensional data of the pantograph and the skateboard are collected by the 3D data acquisition system, and the three-dimensional data of the pantograph and the skateboard are obtained, and the three-dimensional data of the pantograph and the skateboard are processed by the skateboard extraction module and the three-dimensional registration module, wherein the skateboard extraction module includes a skateboard extraction algorithm and a skateboard reference line extraction algorithm, and the three-dimensional registration module includes a deep learning registration algorithm and a 3D precise registration algorithm.

[0055] Figure 3 is the pantograph point cloud depth map, based on Figure 3 The algorithm for extracting the slide plate from the pantograph point cloud depth map is as follows:

[0056] 1) Preprocess the pantograph depth map, that is, use a horizontal rectangular kernel to corrode the depth map. The corrosion process is to convolve the image A with the convolution template B to obtain the minimum value of the pixel points in the area covered by B, and use this minimum value to replace the pixel value of the reference point. In this algorithm, the convolution template uses a horizontal rectangular kernel, which means that the size of the convolution template is (N, 1), that is, the vertical width is 1 and the horizontal length is N (N>1) pixels, which is conducive to corroding the invalid background data in the vertical direction. Then perform the expansion process, that is, convolve the image with the convolution template to obtain the maximum value of the pixel points in the area covered by the template, and use this maximum value to replace the pixel value of the reference point. After the corrosion and expansion processes, the non-slide area can be eliminated to obtain the first pantograph depth map.

[0057] 2) Searching for a connected domain on the first pantograph depth map, that is, Figure 2 Value, traverse the first pantograph depth map, search the 3*3 area of ​​the neighborhood of the non-zero point, if there is a non-zero value in the neighborhood of the non-zero point (that is, a non-zero value in the binary depth map), it is the same connected domain, and then perform a corresponding search on any point in the neighborhood to expand the range of the connected domain, and finally get multiple connected domains, that is, a connected domain containing two skateboards and the corresponding middle connecting axis (the three areas with the largest area contained in the connected domain are the connected domain containing the two skateboards and the middle connecting axis corresponding to the two skateboards), according to the difference in depth values ​​of the skateboards and the connecting axis, remove the middle connecting axis, and you can get a second pantograph depth map containing two skateboards.

[0058] 3) Classifying the second pantograph depth map slide to obtain a third pantograph depth map including the slide depth map.

[0059] Extract the upper and lower edges of the skateboard depth map in the third pantograph depth map, that is, the pixel coordinates of the skateboard edges in the skateboard depth map, and perform polynomial fitting on the upper and lower edges to obtain the upper and lower edge curves of the skateboard in the depth map. Obtain the upper and lower edge positions in the skateboard depth map through the upper and lower edge curves of the skateboard, thereby obtaining the center coordinates of the skateboard (the center coordinates are the midpoints of the upper and lower edge positions); perform linear fitting on the obtained center coordinates of the skateboard to obtain the slope k and intercept b of the reference line, and obtain the linear expression y=kx+b of the skateboard reference line, where (x, y) are the image pixel coordinates. Through this linear expression, the coordinates of the skateboard reference line in the skateboard depth map are obtained. The three-dimensional point cloud data of the skateboard reference line can be obtained through the mapping relationship between the skateboard depth map and the point cloud, and then the skateboard reference line to be tested can be obtained. The position of the obtained point cloud reference line in the pantograph is as follows: Figure 4 shown.

[0060] The method for extracting the reference line of the standard skateboard based on the standard skateboard depth map is the same as the method for extracting the reference line of the skateboard to be tested based on the pantograph depth map, and has been recorded above and will not be repeated here.

[0061] Obtain the standard skateboard reference line point cloud data, randomly rotate (in three directions of X, Y, and Z) and translate (in three directions of X, Y, and Z) the standard skateboard reference line point cloud data, and add noise to the connected domain of the middle connecting axis corresponding to the two skateboards to simulate the skateboard wear point cloud. The noise can be random noise, and then obtain the training set for deep learning registration;

[0062] The data in the training set are normalized and sent to the twin network based on PointNet as the backbone network to obtain two feature vectors (each feature vector is a 1×256 value of the feature value automatically extracted by the network); based on these two feature vectors, the twin network is used to fuse and finally regress six quantities, which represent three rotation quantities and three translation quantities of the X, Y, and Z axes respectively. The MSE loss function is used to perform reverse gradient calculation to update the network parameters to obtain a deep learning registration training network;

[0063] Acquire the point cloud data of the reference line of the skateboard to be tested, input the point cloud data of the standard skateboard reference line and the point cloud data of the reference line of the skateboard to be tested into the deep learning alignment training network, and obtain the first rotation amount and the first translation amount of the reference line of the skateboard to be tested relative to the standard skateboard reference line; calculate the first rotation amount and the first translation amount to obtain the point cloud data of the reference line of the first calibrated skateboard to be tested, and obtain the first calibrated reference line of the skateboard to be tested according to the mapping relationship.

[0064] The registration reference lines obtained by the deep learning registration algorithm and the standard reference lines are in the same plane. However, due to the wear of the skateboard during actual use, there is a certain deviation between the data after deep learning registration and the standard reference lines. More accurate registration is required again. The specific steps are as follows:

[0065] L1: performing filtering and noise reduction processing on the first calibration slide reference line to be tested to obtain the first calibration slide reference line to be tested after noise reduction;

[0066] The point cloud statistical filtering algorithm is used in the filtering and noise reduction process. The point cloud statistical filtering algorithm can be understood as follows: the characteristic of obvious outliers is that they are sparsely distributed in space, and each point expresses a certain amount of information. The denser the points in a certain area, the greater the amount of information may be. Noise information is useless information and has a small amount of information. Therefore, the information expressed by outliers can be ignored. Taking into account the characteristics of outliers, it can be defined that the point cloud is less than a certain density, that is, the point cloud is invalid. Calculate the average distance from each point to its nearest k points, that is, assume that the result is a Gaussian distribution, whose shape is determined by the mean and standard deviation, then points with an average distance outside the standard range can be defined as outliers and removed from the data.

[0067] L2: Calculating the distance between the first calibration test skateboard reference line after noise reduction and the corresponding point of the standard skateboard reference line;

[0068] Setting a threshold, identifying the portion of the distance greater than the threshold as a wear interval, and removing the wear intervals of the first calibration test slide reference line and the standard slide reference line after noise reduction;

[0069] L3: According to formula (1), the first calibration test slide reference line and the standard slide reference line after the wear interval is removed are calculated to obtain the distance from the first calibration test slide reference line after the wear interval is removed to the standard slide reference line after the wear interval is removed.

[0070]

[0071] Among them, x1, y1, z1 represent the standard reference line data, x2, y2, z2 represent the reference line data after deep learning registration, and r 11 -r 33 represents the amount of rotation, t x , t y , t z They represent the translation vectors respectively;

[0072] L4: Iterate L2 and L3. When the distance from the first calibrated skateboard reference line after the wear interval is removed to the standard skateboard reference line after the wear interval is removed is less than a set threshold, stop iteration to obtain the second calibrated skateboard reference line.

[0073] The upper surface curve of the skateboard formed by the reference line of the skateboard to be tested and the standard reference line data obtained by the above algorithm is as follows: Figure 5 As shown in the figure, the horizontal axis is the coordinate system of the slide plate (from the lowest point of the left horn to the lowest point of the right horn), and the vertical axis is the distance from any point on the upper surface of the slide plate to the straight line formed by the lowest point of the left horn and the lowest point of the right horn; the upper curve is the standard reference line, and the lower curve is the reference line of the slide plate to be tested. The wear value of the pantograph slide plate = standard value - slide plate value. And through Figure 5 It can be seen that the wear area of ​​the skateboard to be tested is mainly in the middle area of ​​the skateboard. The wear becomes smaller as it moves to both sides, and there is basically no wear on the horn part. This result is exactly the same as the actual wear of the skateboard.

[0074] It should be noted that the conversion between the depth map and the three-dimensional point cloud data recorded above are all commonly used mapping relationships, and are not the core technical content of this application, and will not be described in detail here.

[0075] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0076] Acquire pantograph three-dimensional data, convert the pantograph three-dimensional data into a pantograph depth map, and extract a reference line of a slide plate to be tested from the pantograph depth map;

[0077] Acquire three-dimensional data of a standard skateboard, convert the three-dimensional data of the standard skateboard into a standard skateboard depth map, and extract a standard skateboard reference line from the standard skateboard depth map;

[0078] The reference line of the slide to be tested is three-dimensionally registered with the reference line of the standard slide to obtain the difference between the reference line of the slide to be tested and the reference line of the standard slide, and the pantograph wear is calculated based on the difference.

[0079] The above embodiment provides a computer device, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0080] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: obtaining pantograph three-dimensional data, converting the pantograph three-dimensional data into a pantograph depth map, and extracting a reference line of a slide plate to be tested from the pantograph depth map;

[0081] Acquire three-dimensional data of a standard skateboard, convert the three-dimensional data of the standard skateboard into a standard skateboard depth map, and extract a standard skateboard reference line from the standard skateboard depth map;

[0082] The reference line of the slide to be tested is three-dimensionally registered with the reference line of the standard slide to obtain the difference between the reference line of the slide to be tested and the reference line of the standard slide, and the pantograph wear is calculated based on the difference.

[0083] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0084] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0085] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for calculating the wear of a pantograph slide plate, characterized in that: The method comprises: Obtain pantograph three-dimensional data, convert the pantograph three-dimensional data into a pantograph depth map, and extract a skateboard depth map from the pantograph depth map according to a skateboard extraction algorithm, the skateboard extraction algorithm comprising: 1) preprocessing the pantograph depth map, eliminating non-skateboard areas, and obtaining a first pantograph depth map; 2) searching for a connected domain on the first pantograph depth map, the three areas with the largest areas being the connected domains containing two skateboards and the middle connecting axes corresponding to the two skateboards; according to the difference in depth values ​​of the skateboards and the connecting axes, removing the middle connecting axis of the connected domain to obtain a second pantograph depth map containing the two skateboards; 3) classifying the skateboards in the second pantograph depth map to obtain a third pantograph depth map containing a skateboard depth map, and extracting a reference line of the skateboard to be tested from the pantograph depth map; Acquire three-dimensional data of a standard skateboard, convert the three-dimensional data of the standard skateboard into a standard skateboard depth map, and extract a standard skateboard reference line from the standard skateboard depth map; The reference line of the slide to be tested is three-dimensionally registered with the reference line of the standard slide to obtain the difference between the reference line of the slide to be tested and the reference line of the standard slide, and the pantograph wear is calculated based on the difference.

2. A method for calculating pantograph slide wear according to claim 1, characterized in that: The method of acquiring the pantograph three-dimensional data, converting the pantograph three-dimensional data into a pantograph depth map, and extracting the reference line of the slide plate to be tested from the pantograph depth map is specifically as follows: Acquire pantograph three-dimensional data, and convert the pantograph three-dimensional data into a pantograph depth map; extracting a slide depth map from the pantograph depth map according to the slide extraction algorithm; The reference line of the skateboard to be tested is extracted from the skateboard depth map according to the skateboard reference line extraction algorithm.

3. The method for calculating the wear of a pantograph slide plate according to claim 1, characterized in that: The step of extracting the reference line of the skateboard to be tested from the skateboard depth map according to the skateboard reference line extraction algorithm is specifically as follows: Extracting upper and lower edges of the slide depth map, performing polynomial fitting on the upper and lower edges, and obtaining upper and lower edge curves of the slide in the slide depth map; Obtaining upper and lower edge positions of the skateboard in the skateboard depth map through upper and lower edge curves of the skateboard in the skateboard depth map, and then obtaining the center coordinates of the skateboard; Performing a straight line fitting on the center coordinates of the skateboard to obtain a straight line expression of the skateboard reference line; The coordinates of the skateboard reference line in the skateboard depth map are obtained through the straight line expression, and the three-dimensional data of the skateboard reference line is obtained through the mapping relationship, thereby obtaining the skateboard reference line to be tested.

4. A method for calculating pantograph slide wear according to claim 3, characterized in that: The acquiring of the standard skateboard three-dimensional data, converting the standard skateboard three-dimensional data into a standard skateboard depth map, and extracting the standard skateboard reference line from the standard skateboard depth map are specifically as follows: Based on the skateboard reference line extraction algorithm, a standard skateboard reference line is extracted from a standard skateboard depth map.

5. The method for calculating the wear of a pantograph slide plate according to claim 3, characterized in that: The three-dimensional registration of the reference line of the slide to be tested and the reference line of the standard slide is performed to obtain the difference between the reference line of the slide to be tested and the reference line of the standard slide, and the pantograph wear is calculated according to the difference as follows: Performing deep learning registration on the reference line of the skateboard to be tested and the reference line of the standard skateboard to obtain a first calibrated reference line of the skateboard to be tested; Performing 3D precise registration on the first calibration test skateboard reference line to obtain a second calibration test skateboard reference line; The pantograph wear is calculated based on the difference between the reference line of the second calibration slide to be tested and the reference line of the standard slide.

6. A method for calculating pantograph slide wear according to claim 5, characterized in that: The deep learning registration of the reference line of the skateboard to be tested and the reference line of the standard skateboard to obtain the first calibrated reference line of the skateboard to be tested is specifically: Obtaining standard skateboard reference line point cloud data, rotating and translating the standard skateboard reference line point cloud data, adding noise to the connected domain of the middle connecting axes corresponding to the two skateboards, and obtaining a deep learning registration training set; According to the deep learning registration training set, the network parameters are updated by using the MSE loss function to perform reverse gradient calculation to obtain a deep learning registration training network; Acquire the point cloud data of the reference line of the skateboard to be tested, input the point cloud data of the reference line of the standard skateboard and the point cloud data of the reference line of the skateboard to be tested into the deep learning registration training network, and obtain the first rotation amount and the first translation amount of the reference line of the skateboard to be tested relative to the standard skateboard reference line; The first rotation amount and the first translation amount are calculated to obtain the point cloud data of the reference line of the first calibrated skateboard to be tested, and the first calibrated skateboard reference line to be tested is obtained according to the mapping relationship.

7. A method for calculating pantograph slide wear according to claim 6, characterized in that: The 3D precise registration of the first calibration test slide reference line to obtain the second calibration test slide reference line is specifically: L1: performing filtering and noise reduction processing on the first calibration slide reference line to be tested to obtain the first calibration slide reference line to be tested after noise reduction; L2: Calculating the distance between the first calibration test skateboard reference line after noise reduction and the corresponding point of the standard skateboard reference line; Setting a threshold, identifying the portion of the distance greater than the threshold as a wear interval, and removing the wear intervals of the first calibration test slide reference line and the standard slide reference line after noise reduction; L3: According to formula (1), the first calibration test slide reference line and the standard slide reference line after the wear interval is removed are calculated to obtain the distance from the first calibration test slide reference line after the wear interval is removed to the standard slide reference line after the wear interval is removed. Among them, x1, y1, z1 represent the standard reference line data, x2, y2, z2 represent the reference line data after deep learning registration, and r 11 -r 33 represents the amount of rotation, t x , t y , t z They represent translation vectors respectively; L4: Iterate L2 and L3. When the distance from the first calibrated skateboard reference line after the wear interval is removed to the standard skateboard reference line after the wear interval is removed is less than a set threshold, stop iteration to obtain the second calibrated skateboard reference line.

8. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images