A method and system for detecting pantograph component dimensions based on 3D images

Abstract

The present invention discloses a method and system for detecting the size of pantograph components based on 3D images, and relates to the technical field of electric locomotive safety monitoring. The detection method includes: S1, collecting a 3D image of the pantograph of an incoming vehicle; S2, preprocessing the 3D image of the pantograph to obtain pantograph point cloud data; S3, performing key size detection of the carbon slide plate and carbon slide plate wear detection based on the pantograph point cloud data; S4, integrating the size detection results and the wear detection results and outputting them. The present invention utilizes 3D image acquisition technology to perform real-time 3D image acquisition of the pantograph components of the vehicle, and realizes the acquisition of multi-size information with a single acquisition source. Compared with two-dimensional data or multi-source data, the data acquisition and processing costs are reduced, and the pantograph can be comprehensively monitored in multiple sizes.

Description

Technical Field

[0001] The present invention relates to the technical field of electric locomotive safety monitoring, and in particular to a method and system for detecting the size of a pantograph component based on a 3D image. Background Art

[0002] The pantograph is a device that electric locomotives use to obtain electricity from the power grid. It is always in contact with the contact wire when the vehicle is running. Long-term contact friction will cause the carbon plate of the pantograph to wear. When the carbon plate is worn below the safe thickness, it may cause the pantograph to separate from the contact wire, leading to safety issues such as offline and temporary power outages. Therefore, how to effectively, frequently and comprehensively monitor the wear status of the pantograph plate is an important guarantee for driving safety.

[0003] In recent years, with the development of machine vision technology and deep learning, many methods based on machine vision technology and deep learning have been applied to the detection of pantograph dimensions. Among them, the most common method is to use a 2D camera to obtain the surface information of the pantograph in the horizontal direction. This method often requires the cooperation of multiple acquisition devices, or combines different sensors to collect data for comprehensive analysis. The fusion processing of multi-source data increases the complexity of the algorithm, and the plane information is insufficient to obtain other information such as height, depth, overall offset and distortion in the vertical direction, resulting in the inability to fully monitor the status of the pantograph. Summary of the invention

[0004] The present invention provides a method and system for detecting the dimensions of pantograph components based on 3D images, so as to solve the problem that the existing pantograph monitoring method is insufficient in acquiring information.

[0005] The present invention is achieved through the following technical solutions:

[0006] A first aspect of the present invention provides a method for detecting the size of a pantograph component based on a 3D image, comprising:

[0007] S1, collects the 3D image of the pantograph of the oncoming vehicle;

[0008] S2, preprocessing the pantograph 3D image to obtain pantograph point cloud data;

[0009] S3, based on the pantograph point cloud data, perform key size detection and carbon slide wear detection to obtain the detection results.

[0010] The present invention utilizes 3D image acquisition technology to perform real-time acquisition of 3D images of the pantograph components of a vehicle. Only a single acquisition device is required to directly obtain the three-dimensional data of the pantograph. By performing multi-scale analysis on the three-dimensional point cloud data, the acquisition of multi-scale information from a single acquisition source is achieved. Compared with two-dimensional data or multi-source data, the data acquisition and processing costs are reduced, and the pantograph can be comprehensively monitored for multi-scale information.

[0011] In one implementation, collecting the pantograph 3D image of the vehicle includes: collecting the pantograph 3D image of the oncoming vehicle by matching a collection frequency with a speed of the oncoming vehicle.

[0012] In one implementation, preprocessing the pantograph 3D image to obtain pantograph point cloud data includes: performing file calibration processing on the pantograph 3D image according to a calibration file to obtain calibrated point cloud data.

[0013] In one embodiment, the preprocessing further includes: if the calibrated point cloud data is dispersed into multiple files, then the point cloud data in the multiple files are spliced ​​to obtain complete point cloud data.

[0014] In one implementation, the step S1 specifically includes: the 3D image acquisition device acquires the 3D image of the pantograph of the oncoming vehicle in response to the image acquisition signal, and stops the acquisition action in response to the stop acquisition signal.

[0015] In one embodiment, the carbon slide plate key dimension detection and carbon slide plate wear detection based on the pantograph point cloud data includes:

[0016] S301, positioning key points of the pantograph point cloud data according to the size characteristics of the pantograph component to obtain carbon slide plate point cloud data;

[0017] S302, performing horizontal correction on the carbon skateboard point cloud data and establishing a three-dimensional coordinate system;

[0018] S303, in the three-dimensional coordinate system, performing carbon slide board key dimension detection and carbon slide board wear detection based on the carbon slide board point cloud data, wherein the key dimensions include the carbon slide board width, the spacing between two carbon slide boards, and the deflection angle.

[0019] In one implementation, step S301 is specifically as follows:

[0020] S3011, downsampling the pantograph point cloud data, removing invalid point cloud data that does not belong to the pantograph by high-level screening, performing connectivity analysis and point cloud quantity screening on the remaining point cloud data, and removing noise point clouds;

[0021] S3012, triangulating the point cloud data processed in step S3011, and locating the key position of the carbon slide board using the height and length features of the carbon slide board;

[0022] S3013, using the key position of the carbon slide to crop the pantograph point cloud data, and then using the normal vector to remove the edge interference points of the carbon slide to obtain the carbon slide point cloud data.

[0023] In one embodiment, the carbon slide wear detection includes:

[0024] S3041, for each carbon slide point cloud set, taking a longitudinal mid-section of the carbon slide, and converting the section into a contour line;

[0025] S3042, leveling the contour line;

[0026] S3043, in the three-dimensional coordinate system, detecting the highest points H0 and H1 at both ends of the contour line, connecting H0 and H1 to obtain a reference line, and determining the area between the reference line and the contour line as the wear area;

[0027] S3044, in the wear area, the distance from the contour line to the reference line is measured at intervals of 1 mm to obtain the wear curve of the carbon slide plate;

[0028] S3045, obtaining the maximum wear value and its corresponding position according to the wear curve.

[0029] In one embodiment, the method further includes: performing three-dimensional modeling of the pantograph carbon skateboard based on the size detection results and the wear detection results to obtain a carbon skateboard rendering, and displaying the carbon skateboard rendering through a business end interface.

[0030] A second aspect of the present invention provides a pantograph component size detection system based on 3D images, comprising:

[0031] A 3D image acquisition module is configured to acquire a 3D image of the pantograph of an oncoming vehicle;

[0032] A data processing module is configured to pre-process the pantograph 3D image to obtain pantograph point cloud data;

[0033] A data analysis module is configured to perform a carbon slide plate key dimension detection and a carbon slide plate wear detection based on the pantograph point cloud data;

[0034] The output module is configured to integrate the size detection result and the wear detection result and output them.

[0035] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0036] 3D image acquisition technology is used to collect 3D images of the pantograph components of the vehicle in real time, realizing the acquisition of multi-size information with a single acquisition source. Compared with two-dimensional data or multi-source data, it reduces the cost of data acquisition and processing, and can perform comprehensive multi-size monitoring of the pantograph.

[0037] The 3D image acquisition is triggered by the acquisition signal, and only the pantograph components are acquired to avoid the introduction of irrelevant images.

[0038] By matching the acquisition frequency of the vehicle speed to collect data on the pantograph components, stretching or compression of the collected data due to different vehicle speeds can be avoided, thereby ensuring the quality of the original collected data.

[0039] The data collected by the pantograph of the oncoming vehicle is calibrated using the vehicle number information to avoid data confusion.

[0040] Point cloud data processing is based on the same reference coordinate system to improve algorithm accuracy.

[0041] It provides accurate detection results and visualized 3D reconstruction images, provides timely and effective maintenance guidance for maintenance personnel, and provides more reliable protection for running trains. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0043] Figure 1 is a flow chart of a method for detecting the size of a pantograph component based on a 3D image according to an embodiment of the present invention;

[0044] Figure 2 This is a 3D data collection effect diagram of a pantograph according to an embodiment of the present invention;

[0045] Figure 3 A three-dimensional coordinate system of pantograph point cloud data according to an embodiment of the present invention;

[0046] Figure 4 This is a schematic diagram of calculating the wear amount of a carbon slide plate according to an embodiment of the present invention;

[0047] Figure 5 This is the hardware structure of a pantograph component size detection system based on 3D images according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0049] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to or inherent to other steps or units of the device.

[0050] The terms used in the various embodiments of the application are only used to describe the purpose of specific embodiments and are not intended to limit the various embodiments of the application. As used herein, the singular form is intended to also include the plural form, unless the context clearly indicates otherwise. Unless otherwise limited, all terms used here (including technical terms and scientific terms) have the same meaning as the meanings commonly understood by ordinary technicians in the field of the various embodiments of the application. The terms (such as the terms defined in the dictionary generally used) will be interpreted as having the same meaning as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized meaning or an overly formal meaning, unless clearly defined in the various embodiments of the application.

[0051] The embodiment of the present invention provides a method for detecting the size of a pantograph component based on a 3D image, which is suitable for real-time monitoring of the size of a pantograph component of a vehicle and is conducive to comprehensive monitoring of pantographs in multiple sizes.

[0052] like Figure 1 As shown, Figure 1 The present invention is a flow chart of a method for detecting the size of a pantograph component based on a 3D image, the method comprising the following steps:

[0053] S1, collects the 3D image of the pantograph of the oncoming vehicle;

[0054] S2, preprocessing the pantograph 3D image to obtain pantograph point cloud data;

[0055] S3, based on the pantograph point cloud data, the key dimensions of the carbon slide plate and the wear amount of the carbon slide plate are detected to obtain the test results.

[0056] The present invention aims to capture the image of the pantograph through a 3D image acquisition device when the pantograph passes through the detection area without affecting the normal operation of the vehicle, and reconstruct the shape and geometric features of the pantograph in three-dimensional space. Through background software processing and calculation, the surface information detection results of the pantograph carbon slide plate, such as the wear amount, carbon slide plate width, carbon slide plate spacing, and deflection angle, can be obtained.

[0057] The 3D image acquisition device uses a 3D digital camera (hereinafter referred to as 3D camera), which can realize the three-dimensional data acquisition of images and videos. The 3D camera is installed in the center of the electric track, with the shooting direction vertically downward and perpendicular to the driving direction of the vehicle. It is used to take the 3D image of the pantograph from a bird's-eye view, and obtain the pantograph image information in the center of the acquisition field of view, covering the entire pantograph. Figure 2 Shown is the effect diagram of pantograph 3D data collection.

[0058] In step S1, a 3D image of a pantograph of a moving vehicle is collected by a 3D camera, specifically including: the 3D camera collects the 3D image of the pantograph of the moving vehicle in response to a received image collection signal, and stops collecting in response to a stop collection signal. In this way, only the pantograph components are collected, and less irrelevant images are introduced, thereby improving the accuracy of the collected data and reducing the complexity of subsequent processing.

[0059] The image acquisition signal and the stop acquisition signal are controlled by the pantograph positioning device, which consists of an infrared beam group A and an infrared beam group B fixed on the track and spaced a certain distance apart. When the pantograph passes through the infrared beam group A, the pantograph positioning device sends a detection signal to the main control device, which generates an image acquisition signal based on the signal and sends it to the 3D camera. The 3D camera starts the acquisition action after receiving the signal. When the pantograph passes through the infrared beam group 2, the pantograph positioning device sends another detection signal to the main control device, which generates a stop acquisition signal based on the signal and sends it to the 3D camera. The 3D camera stops the acquisition action after receiving the signal. During the period of receiving the image acquisition signal and the stop acquisition signal, the 3D camera acquires multiple pantograph images at a predetermined acquisition frequency to ensure that sufficient data is acquired.

[0060] In the direction of vehicle travel, the vehicle first reaches infrared beam group A, and then reaches infrared beam group B, ensuring that the 3D image acquisition device acquires the complete pantograph components. After acquisition, the acquired data is transmitted to the main control device via a wired or wireless method for subsequent processing. The above-mentioned main control device is an intelligent terminal and / or server with control signal transmission function and data processing function, which can be a single terminal or server, or a combination of multiple terminals and / or multiple servers.

[0061] In a preferred embodiment, it is set that within the time of 2S, the infrared radiation group A and the infrared radiation group B are effectively triggered only once to avoid sending wrong signals due to sensing errors and affecting the normal operation of the 3D image acquisition device. It should be noted that "2s" is only used as an example, and the specific time can be determined according to the time it takes for the pantograph to completely pass through the detection area.

[0062] In one embodiment, before step S1, a step of acquiring vehicle number information of an incoming vehicle is further included, specifically: the vehicle number collection device collects vehicle number information of a passing vehicle in response to a vehicle number collection signal.

[0063] The vehicle number collection signal is controlled by the vehicle number magnetic steel sensor installed on the track. In the direction of vehicle travel, the vehicle first passes the vehicle number magnetic steel sensor and then passes the infrared beam group. When the vehicle passes over the vehicle number magnetic steel sensor, the vehicle number collection signal is triggered and the main control device starts the vehicle number collection camera to collect the vehicle number. After the collection is completed, the vehicle number information is analyzed.

[0064] The calibration file is determined according to the vehicle number information, and the pantograph 3D image is calibrated according to the calibration file to obtain the calibrated point cloud data.

[0065] The 3D camera collects data from the pantograph multiple times between the time when it receives the image acquisition signal and the time when it ends the acquisition signal. After the acquisition is completed, the multiple 3D images are transmitted to the back-end device for processing, and the data from different vehicles are distinguished according to the vehicle number information.

[0066] Furthermore, the step of preprocessing the 3D image includes: performing file calibration processing on the pantograph 3D image according to the calibration file to obtain calibrated point cloud data. The calibration file is preset, and the calibrated data is point cloud data or depth image in TXT format.

[0067] Furthermore, the preprocessing further includes: if the calibrated point cloud data is dispersed into multiple files, then the point cloud data in the multiple files are spliced ​​to obtain complete point cloud data. In one embodiment, the method for collecting a 3D image of a pantograph of a moving vehicle includes: collecting a 3D image of a pantograph of an oncoming vehicle by matching an acquisition frequency with an oncoming vehicle speed.

[0068] The speed measuring device completes the speed detection of the oncoming vehicle in response to the oncoming vehicle signal, matches the acquisition frequency corresponding to the vehicle speed, sends the sampling frequency and image acquisition signal to the 3D camera, starts the 3D camera to collect at the acquisition frequency, until the pantograph triggers the stop acquisition sensor, the system completes the acquisition of a single pantograph, and stores the collected data in the set directory. The above-mentioned speed measuring device is a radar speed measuring device or a speed measuring magnet.

[0069] By matching the acquisition frequency of the vehicle speed to collect data on the pantograph components, high-quality data can be collected for different vehicles to avoid stretching or compression of the collected data due to different vehicle speeds. The 3D camera trigger frequency is calculated according to the vehicle speed to drive the camera hard trigger, thereby ensuring the quality of the original collected data.

[0070] For example, if the image resolution is to reach P (mm / piexl), when the measured speed of the oncoming vehicle is S (km / h) = S / 3.6 (m / s) = S / 3.6 (mm / ms), the sampling frequency required for the 3D camera is S / (3.6*P) (KHz), which is the trigger frequency of the camera.

[0071] The vehicle oncoming signal is controlled by the magnetic steel sensor installed on the track. In the direction of vehicle travel, the vehicle first passes the magnetic steel sensor, then the vehicle number magnetic steel sensor and the infrared beam group. When the vehicle presses the magnetic steel sensor, the vehicle oncoming signal is triggered, and the value block device starts the speed measuring device to detect the speed of the vehicle oncoming.

[0072] In one embodiment, it also includes: performing three-dimensional modeling of the pantograph carbon skateboard based on the size detection results and the wear detection results, obtaining a carbon skateboard rendering, and displaying the carbon skateboard rendering through a business end interface.

[0073] Furthermore, it also includes: reconstructing the pantograph point cloud data processed in step S2 in three dimensions to obtain a pantograph rendering and displaying it through a business end interface. The pantograph rendering and the carbon slide rendering can be displayed separately through window switching, or can be displayed side by side in the business end interface.

[0074] Furthermore, the business-end interface displays the test results of size and wear, and reminds personnel to review and handle abnormal values ​​that exceed the threshold through a graded alarm system.

[0075] The graded alarm system sets different levels of alarms, and different levels correspond to different set thresholds. For example, red, yellow, and blue warning levels are set according to the amount of wear. When the amount of wear reaches the wear threshold set for the red level alarm, the graded alarm system reminds the staff to compound the abnormal values ​​with the prompt information corresponding to the red level alarm.

[0076] This method can realize comprehensive monitoring of pantograph components and provide accurate detection results and visual 3D reconstruction images. Such a monitoring system can effectively improve the safety of train operation, provide maintenance personnel with timely and effective maintenance guidance, and provide more reliable protection for running trains.

[0077] Furthermore, after completing a complete pantograph monitoring process for a moving vehicle, the system is initialized and is on standby at any time, waiting for the next oncoming vehicle signal.

[0078] A complete pantograph monitoring process includes the process from the incoming vehicle magnetic steel sensor detecting the incoming vehicle signal to the outgoing vehicle magnetic steel sensor detecting the outgoing vehicle signal. In the direction of vehicle travel, the vehicle passes through the incoming vehicle magnetic steel sensor, the vehicle number magnetic steel sensor, the infrared beam group A, the infrared beam group B and the outgoing vehicle magnetic steel sensor in sequence. When the wheel passes over the outgoing vehicle magnetic steel sensor, the outgoing vehicle magnetic steel starts counting. When the outgoing vehicle magnetic steel count is (4M-1) or 4M, a outgoing vehicle signal is generated after a delay of several seconds (for example, 2s), where M is the number of train formations or the number of wheel groups. The outgoing vehicle signal is generated after a delay of several seconds to ensure that the vehicle has completely passed the monitoring area.

[0079] In one embodiment, when the vehicle triggers the departure signal, the departure signal is broadcast to the data processing module, and the data processing module pre-processes the collected data received from the 3D camera in the current monitoring cycle, including file calibration, file name verification, etc., and then transmits the pre-processed data to the pantograph analysis module for size detection and carbon slide wear detection. The processing of the collected data is triggered after the vehicle leaves the detection area to avoid data loss due to network delays and other reasons, and ensure the accuracy of data processing.

[0080] In step S3, the pantograph point cloud data obtained in step S2 is processed to obtain the key dimension data and the wear data of the carbon slide plate, including the following steps:

[0081] S301, positioning key points of the pantograph point cloud data according to the size characteristics of the pantograph components to obtain carbon slide plate point cloud data.

[0082] S302, performing horizontal correction on the carbon skateboard point cloud data and establishing a three-dimensional coordinate system.

[0083] S303, in a three-dimensional coordinate system, performing a carbon slide board key dimension detection and a carbon slide board wear detection based on the carbon slide board point cloud data, wherein the key dimensions include the carbon slide board width, the spacing between two carbon slide boards, and the deflection angle.

[0084] After the above tests are completed, the size test results and the wear test results are integrated into the test information and output. The output results are transmitted in the form of messages.

[0085] In step 301, if the power receiving point cloud data is dispersed into multiple files, the point cloud files are first read one by one, and they are spliced ​​to obtain complete point cloud data, and then steps S301-S303 are executed.

[0086] The above step S301 is specifically as follows:

[0087] S3011, downsample the pantograph point cloud data, remove invalid point cloud data that does not belong to the pantograph through height screening, perform connectivity analysis and point cloud quantity screening on the remaining point cloud data, and remove noise point clouds.

[0088] The pantograph point cloud data C1 is downsampled to obtain point cloud data C2, which greatly reduces the amount of data without changing the point cloud distribution, which is sufficient to meet the needs of key point positioning. According to the pantograph height or rail height, the point cloud data below the height, such as the vehicle body, contact wire / net, pantograph upper / lower arm, etc., are filtered to obtain point cloud data C3. The remaining point cloud data are subjected to connectivity analysis and point quantity screening to remove invalid noise points formed by clutter signals such as diffuse reflection near the pantograph to obtain point cloud data C4. Point cloud data C4 retains the main point cloud data of the pantograph, thereby greatly reducing the subsequent calculation time.

[0089] S3012, triangulating the point cloud data processed in step S3011, and locating the key position of the carbon slide board using the height and length features of the carbon slide board.

[0090] Point cloud triangulation is to project the ordered point cloud into the local two-dimensional coordinate plane, link the relationship between each fixed point, triangulate in the plane coordinates, and then establish a triangular surface mesh according to the topological relationship. The point cloud data C4 is triangulated and reconstructed into a surface. Since there are multiple parts on the pantograph, such as the bow angle, the height and length characteristics of the carbon slide are used to locate the key position of the carbon slide, and these noise point clouds are removed. Specifically, a plane that intersects with the carbon slide in the width direction of the carbon slide is established to obtain the intersection with the point cloud, and the height and length characteristics of the carbon slide are used to screen out the position of the carbon slide.

[0091] S3013, using the key position of the carbon slide plate to crop the pantograph point cloud data, and then using the normal vector to remove the interference points at the edge of the carbon slide plate to obtain the carbon slide plate point cloud data.

[0092] The key position is obtained by analyzing the pantograph point cloud data after downsampling. The positioning information is applied to the pantograph point cloud data before downsampling, and the point cloud data belonging to the pantograph carbon slide is obtained by cropping. In the point cloud, each point cloud calculates the normal vector through the surrounding adjacent point cloud information. The point cloud of the carbon slide is almost on the same plane, so the normal vector can be used to identify the carbon slide target and obtain the carbon slide point cloud data. Then, the carbon slide size detection and wear detection are performed based on the point cloud data, and accurate detection results can be obtained.

[0093] In step S302, illustratively, Figure 3It is the three-dimensional coordinate system of the pantograph point cloud data. The three-dimensional coordinate system is established as follows: the x-axis is along the length direction of the pantograph, the y-axis is along the width direction of the pantograph, the planes of the x-axis and y-axis coincide with the upper surface of the pantograph, and the z-axis is vertically downward along the thickness direction of the pantograph. The following calculations are based on this coordinate system, but it can be understood that the description of the coordinate system is only for the convenience of explaining the calculation process. The relative position relationship of the pantograph point cloud data in space does not change due to the coordinate system, and this example is not intended to limit the coordinate system.

[0094] A single device is used for data acquisition, and there is no need for image data stitching. The data is based on the same coordinate system reference, which improves the accuracy of the algorithm.

[0095] In the above step S303, the method for calculating the critical dimensions of the carbon slide plate is specifically as follows:

[0096] S3031, in the xoy plane, at the symmetrical positions on both sides of the x-axis and at a preset distance from the x-axis, the preset distance is the width of the two carbon slides or the spacing between the two rails, usually 400mm, and the edge of the carbon slide is sampled at intervals of 1mm to obtain the edge point sets K1 and K2 of the two carbon slides of the pantograph, such as Figure 3 shown.

[0097] S3032. Based on the edge cloud point cloud sets K1 and K2 obtained above, the width ΔW1 and ΔW2 of each carbon slide board are calculated according to the distance of the edge point clouds. The width can be the average width of the edge point clouds K1 and K2.

[0098] S3033, calculating the deflection angles ΔR1 and ΔR2 of the carbon slide. For each set of edge point clouds of the carbon slide, the least square method is used to perform straight line fitting, and straight lines L1 and L2 are fitted respectively, so as to obtain the deflection angles ΔR1 and ΔR2 of the carbon slide.

[0099] S3034. According to the shortest distance between the edge point clouds of the adjacent sides (inner sides) of the two carbon sliding plates, the average value of each shortest distance is calculated to obtain the pantograph inner plate spacing Δd.

[0100] In the above step S303, the carbon slide wear calculation diagram is as follows: Figure 4 As shown, the specific steps include:

[0101] S3041, for each carbon slide point cloud set, take the longitudinal mid-section of the carbon slide, and convert the section into a contour line.

[0102] S3042, level the contour line. Due to uneven force on the left and right sides of the pantograph, the carbon slide plate may tilt up and down.

[0103] S3043, in the three-dimensional coordinate system, detect the highest points H0 and H1 at both ends of the contour line, connect H0 and H1 to obtain a baseline, and determine the area between the baseline and the contour line as the wear area.

[0104] S3044, in the wear area, the distance from the contour line to the reference line is measured at intervals of 1 mm to obtain the wear curve of the carbon slide plate;

[0105] S3045. Obtain the maximum wear value and its corresponding position according to the wear curve.

[0106] A second aspect of the present invention provides a pantograph component size detection system based on 3D images, comprising:

[0107] A 3D image acquisition module is configured to acquire a 3D image of the pantograph of an oncoming vehicle;

[0108] A data processing module is configured to pre-process the pantograph 3D image to obtain pantograph point cloud data;

[0109] A data analysis module is configured to perform a key dimension detection of the carbon slide plate and a wear detection of the carbon slide plate based on the pantograph point cloud data to obtain a detection result;

[0110] Furthermore, the above system also includes an output module configured to integrate and output the size detection result and the wear detection result.

[0111] Furthermore, the 3D image acquisition module acquires the pantograph 3D image of the oncoming vehicle by matching the acquisition frequency of the oncoming vehicle speed.

[0112] Furthermore, the above system also includes an oncoming vehicle detection module, which is used to obtain the vehicle license plate information of the oncoming vehicle before collecting the pantograph 3D image of the oncoming vehicle.

[0113] Furthermore, the data processing module is also used to determine the calibration file according to the vehicle number information, and perform file calibration processing on the pantograph 3D image according to the calibration file to obtain calibrated point cloud data.

[0114] Furthermore, the data processing module is also used for: if the calibrated point cloud data is scattered into multiple files, then the point cloud data in the multiple files are spliced ​​to obtain complete point cloud data.

[0115] Furthermore, the 3D image acquisition device acquires the 3D image of the pantograph of the oncoming vehicle in response to the image acquisition signal, and stops the acquisition action in response to the stop acquisition signal.

[0116] Furthermore, the data analysis module performs carbon slide plate key dimension detection and carbon slide plate wear detection based on the pantograph point cloud data, specifically including:

[0117] S301, positioning key points of the pantograph point cloud data according to the size characteristics of the pantograph component to obtain carbon slide plate point cloud data;

[0118] S302, performing horizontal correction on the carbon skateboard point cloud data and establishing a three-dimensional coordinate system;

[0119] S303, in the three-dimensional coordinate system, performing carbon slide board key dimension detection and carbon slide board wear detection based on the carbon slide board point cloud data, wherein the key dimensions include the carbon slide board width, the spacing between two carbon slide boards, and the deflection angle.

[0120] Furthermore, the data analysis module is also used to perform three-dimensional modeling of the pantograph carbon slide board based on the size detection results and the wear detection results to obtain the carbon slide board rendering.

[0121] Furthermore, the above system also includes a business end interface for displaying the carbon skateboard renderings and / or detection results through the business end interface.

[0122] The hardware structure of the pantograph component size detection system based on 3D images of the present invention is as follows: Figure 5 As shown, it includes three parts: trackside detection equipment, machine room and main control room. Among them, the trackside detection equipment includes 3D detection components, magnetic steel sensors, speed sensors, pantograph positioning devices and car number cameras. Each trackside detection equipment communicates with the system machine room through the pantograph electric control cabinet, and the system machine room interacts with the main control room. The industrial computer controls the trackside detection equipment to collect signals, receives the collected data for processing and analysis, and transmits the results to the main control room. The main control room issues an alarm through the sound and light graded alarm device according to the detection results.

[0123] The following is a hardware description of the execution process of the pantograph component size detection method based on 3D images of the present invention, which includes the following steps:

[0124] S10, the system completes initialization.

[0125] S21, the oncoming vehicle presses the oncoming vehicle magnetic steel sensor, triggering an oncoming vehicle signal. The pantograph control cabinet receives the signal and controls the speed sensor to detect the speed of the oncoming vehicle.

[0126] S30: The speed sensor completes the speed measurement, and the approaching vehicle passes over the vehicle number magnetic steel sensor, triggering the vehicle number collection signal. The signal is received by the pantograph electric control box, and the electric control box and the main control box CAN communication forward the signal to the industrial computer. The industrial computer 1 starts the vehicle number camera and notifies the host software to enter the collection preparation state.

[0127] S40: The vehicle number collection is completed and the vehicle number information is analyzed, and the vehicle number information is pushed to the main server.

[0128] S50: When the pantograph of the oncoming vehicle passes through infrared beam group 1#, only one effective trigger is performed within 2S, and the electric control cabinet notifies the 3D detection component to start collecting.

[0129] S60: When the pantograph of the oncoming vehicle passes through infrared beam group 2#, only one effective trigger is performed within 2S. At this time, the electric control cabinet notifies the 3D detection component to end the collection, and the collected data is stored on industrial computer 2 through the network cable.

[0130] S60: When the wheels of the oncoming train press over the departure magnet, the departure magnet starts counting. When the departure magnet count reaches 4M-1 (M is the number of train formations), there is a delay of 2 seconds, or when the departure magnet count reaches 4M (M is the number of train formations), a departure signal is generated, the file name of the collected image is verified, and then the departure signal is immediately broadcast to the 3D pantograph analysis module.

[0131] S70: After receiving the task signal, the pantograph analysis module starts to perform calculations. After the calculations are completed, the data is transmitted to the data storage server, and the pantograph analysis module enters a standby state.

[0132] S80: The display terminal gives the latest data of the signal vehicle and the information of the faulty vehicle, completing the whole process of receiving the vehicle, and the system enters the standby state.

[0133] The detection system adopts a multi-process mode, which consists of multiple processes and interacts through agreed methods and indirect communication methods, so that the entire system has flexible and scalable distributed functions and modularity.

[0134] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for detecting the dimensions of pantograph components based on 3D images, characterized in that: include: S1, collects the 3D image of the pantograph of the oncoming vehicle; S2, preprocessing the pantograph 3D image to obtain pantograph point cloud data; S3, based on the pantograph point cloud data, performing a carbon slide plate key dimension detection and a carbon slide plate wear detection to obtain a detection result, including: S3011, downsampling the pantograph point cloud data, removing invalid point cloud data that does not belong to the pantograph by high-level screening, performing connectivity analysis and point cloud quantity screening on the remaining point cloud data, and removing noise point clouds; S3012, triangulating the point cloud data processed in step S3011, and locating the key position of the carbon slide board using the height and length features of the carbon slide board; S3013, using the key position of the carbon slide to crop the pantograph point cloud data, and then using the normal vector to remove the edge interference points of the carbon slide to obtain the carbon slide point cloud data S3014, performing horizontal correction on the carbon skateboard point cloud data and establishing a three-dimensional coordinate system; S3015, in the three-dimensional coordinate system, performing carbon slide board key dimension detection and carbon slide board wear detection based on the carbon slide board point cloud data, wherein the key dimensions include the carbon slide board width, the spacing between two carbon slide boards, and the deflection angle.

2. The method for detecting pantograph components size based on 3D images according to claim 1, characterized in that: The collecting of the pantograph 3D image of the oncoming vehicle includes: collecting the pantograph 3D image of the oncoming vehicle by matching the collecting frequency with the speed of the oncoming vehicle.

3. The method for detecting pantograph components size based on 3D images according to claim 1, characterized in that: The pantograph 3D image is preprocessed to obtain pantograph point cloud data, including: performing file calibration processing on the pantograph 3D image according to a calibration file to obtain calibrated point cloud data.

4. The method for detecting pantograph components size based on 3D images according to claim 3, characterized in that: The preprocessing further includes: if the calibrated point cloud data is dispersed into multiple files, then the point cloud data in the multiple files are spliced ​​to obtain complete point cloud data.

5. The method for detecting pantograph components size based on 3D images according to claim 1, characterized in that: Step S1 specifically includes: the 3D image acquisition device acquires the pantograph 3D image of the oncoming vehicle in response to the image acquisition signal, and stops the acquisition action in response to the stop acquisition signal.

6. The method for detecting pantograph components size based on 3D images according to claim 1, characterized in that: The carbon slide wear detection includes: S3041, for each carbon slide point cloud set, taking a longitudinal mid-section of the carbon slide, and converting the section into a contour line; S3042, leveling the contour line; S3043, in the three-dimensional coordinate system, detecting the highest points H0 and H1 at both ends of the contour line, connecting H0 and H1 to obtain a reference line, and determining the area between the reference line and the contour line as the wear area; S3044, in the wear area, the distance from the contour line to the reference line is measured at intervals of 1 mm to obtain the wear curve of the carbon slide plate; S3045, obtaining the maximum wear value and its corresponding position according to the wear curve.

7. The method for detecting pantograph components size based on 3D images according to claim 1, characterized in that: The method also includes: performing three-dimensional modeling of the pantograph carbon skateboard based on the size detection results and the wear detection results to obtain a carbon skateboard rendering, and displaying the carbon skateboard rendering through a business end interface.

8. A pantograph component size detection system based on 3D images, characterized in that: include: A 3D image acquisition module is configured to acquire a 3D image of the pantograph of an oncoming vehicle; A data processing module is configured to pre-process the pantograph 3D image to obtain pantograph point cloud data; The data analysis module is configured to perform carbon slide plate key dimension detection and carbon slide plate wear detection based on the pantograph point cloud data to obtain detection results, including: S3011, downsampling the pantograph point cloud data, removing invalid point cloud data that does not belong to the pantograph by high-level screening, performing connectivity analysis and point cloud quantity screening on the remaining point cloud data, and removing noise point clouds; S3012, triangulating the point cloud data processed in step S3011, and locating the key position of the carbon slide board using the height and length features of the carbon slide board; S3013, using the key position of the carbon slide to crop the pantograph point cloud data, and then using the normal vector to remove the edge interference points of the carbon slide to obtain the carbon slide point cloud data S3014, performing horizontal correction on the carbon skateboard point cloud data and establishing a three-dimensional coordinate system; S3015, in the three-dimensional coordinate system, performing carbon slide board key dimension detection and carbon slide board wear detection based on the carbon slide board point cloud data, wherein the key dimensions include the carbon slide board width, the spacing between two carbon slide boards, and the deflection angle.

Images