Method for fast automatic measurement of gauge of trackside equipment

Abstract

The application provides a kind of trackside equipment limit fast automatic measurement method, its device includes measuring car, laser scanner, camera, lens, encoder and operation device;Wherein laser scanner, camera and encoder are set on measuring car, lens is set on camera;Laser scanner, camera and encoder are respectively connected with operation device communication;Image is obtained by setting camera, and target equipment area in image is identified by pre-trained model, whether it is the best image is judged according to the position of frame selection area in image, for the best image, the corresponding mileage data is recorded as the actual mileage of corresponding target equipment, after completing the mileage calculation of all trackside equipment, the limit distance of adjacent trackside equipment is calculated, to judge whether it meets engineering requirements.

Description

Technical Field

[0001] This invention relates to the field of rail transit, and in particular to a method for rapid and automatic measurement of trackside equipment clearance. Background Technology

[0002] In modern life, daily travel is becoming increasingly convenient. Within cities, one can choose public transportation such as buses and subways, while between cities, one can choose trains and high-speed rail. For trams like subways and high-speed rail, track lines need to be laid in advance, and trackside equipment needs to be installed. Trackside equipment generally refers to devices installed at both ends of the track, including signal lights and speedometers. After the trackside equipment is installed, it needs to be inspected by the railway department. This includes verifying whether the installation positions of each device are consistent with the design drawings, and checking the clearance distances between each device to confirm whether the actual installation distances match the pre-construction plan. Currently, trackside equipment inspection is mainly carried out through on-site personnel surveys. However, due to the long track lines and the large amount of data to be measured, the workload for surveyors is heavy, time-consuming, and labor-intensive. Furthermore, because there are many indicators to be tested, dozens of testing tools are required, which are extremely inconvenient to carry. Manual measurement is also inconvenient, time-consuming, and prone to data omissions or measurement errors. Therefore, a simple and quick device and method for measuring the clearances of trackside equipment is needed. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for rapid and automatic measurement of trackside equipment clearance, which is simple in structure and easy to use.

[0004] An automatic trackside equipment clearance measurement device includes a measuring vehicle, a laser scanner, a camera, a lens, an encoder, and a computing device; wherein the laser scanner, camera, and encoder are all mounted on the measuring vehicle, and the lens is mounted on the camera; the laser scanner, camera, and encoder are each communicatively connected to the computing device.

[0005] Furthermore, the laser scanner is positioned in the middle of the measuring vehicle, with its scanning direction being both sides of the measuring vehicle; the scanning direction of the laser scanner is always perpendicular to the direction of travel of the measuring vehicle; the encoder is mounted on the wheels of the measuring vehicle; and the computing device is a service terminal.

[0006] A method for rapid and automatic measurement of trackside equipment clearance includes the following steps:

[0007] Step 1: The measuring vehicle receives the start signal and moves along the track. At the same time, the camera collects image information on both sides of the track, the laser scanner collects scanning data, and the encoder collects encoding data. The collected image data, scanning data, and encoding data are transmitted to the computing device in real time.

[0008] Step 2: The computing device receives image data and reads the simultaneously received encoded data, converts the encoded data into mileage data, and saves the image data and mileage data together; The computing device receives scan data and reads the simultaneously received encoded data, converts the encoded data into mileage data, and saves the scan data and mileage data together.

[0009] Step 3: The measuring vehicle receives a stop signal and stops moving forward on the track; the computing device begins to process the saved image data, scan data, and mileage data;

[0010] Step 4: The computing device completes image detection based on the received image data and corresponding mileage data;

[0011] Step 5: After image detection is completed, the computing device processes the scan data based on the received scan data and the corresponding mileage data to obtain the distance data from the trackside equipment to the track.

[0012] Step 6: Output the information of the trackside equipment and end the process.

[0013] Furthermore, the image detection in step 4 includes the calculation of the clearance distance between trackside devices, which includes the following steps:

[0014] Step 41: The model trained based on deep learning receives the input image data, traverses the image data, reads each frame of the image in sequence, and identifies and selects the target device in each frame of the image.

[0015] Step 42: Read the selected area in each frame of the image, and determine whether it is the best image based on the distance between the center point of the selected area and the corresponding center point of the image; determine the mileage corresponding to the target device based on the mileage data corresponding to the best image.

[0016] Step 43: Obtain the mileage corresponding to all track equipment, sort the track equipment according to the mileage, and calculate the clearance distance between adjacent equipment in terms of mileage.

[0017] Step 44: Organize the boundary distances and output them in a table to complete the process.

[0018] Furthermore, in step 41, the deep learning-based model adopts the YOLOv3 Darknet53 basic network model.

[0019] Furthermore, the selection of the optimal image in step 42 includes the following steps:

[0020] Step 421: Read the selected area in the image and obtain the coordinates of two vertices on any diagonal of the selected area, denoted as (Sx, Sy) and (Ex, Ey) respectively; when reading the selected area of ​​the image for the first time, the selected area of ​​the first frame image is read.

[0021] Step 422: Calculate the coordinates of the midpoint of the two vertices (Sx, Sy) and (Ex, Ey), and represent the coordinates of the midpoint as (Cx, Cy).

[0022] Step 423: Obtain the coordinates of the center point of the image, and based on the coordinates of the midpoint of the selected area and the coordinates of the center point of the image, obtain the horizontal coordinate spacing DistanceCW between them;

[0023] Step 424: Determine whether to traverse all images; if so, proceed to step 425; otherwise, return to step 421.

[0024] Step 425: Traverse all images, select the frame with the smallest horizontal coordinate spacing DistanceCW as the best image, and end the step.

[0025] Furthermore, the image detection in step 4 also includes verification of the trackside equipment installation location, which includes the following steps:

[0026] Step 41a: Obtain the trackside equipment and corresponding mileage data identified during the calculation of clearance distance; obtain the preset trackside equipment installation data table;

[0027] Step 42a: Number the trackside equipment according to the mileage data, and read the ID of the trackside equipment and the corresponding mileage data L1 in sequence; read the ID of the trackside equipment and the corresponding mileage data L2 in the table according to the serial number of the installation data table; the ID of the trackside equipment includes signal lights and speed measuring instruments.

[0028] Step 43a: Determine whether the trackside device ID obtained by identification matches the trackside device ID in the corresponding order in the installation data table; if the IDs match, write "Yes" in the corresponding position in the installation data table; otherwise, write "No".

[0029] Step 44a: Calculate the absolute value of the difference between the mileage data L1 obtained from the identification calculation and the mileage data L2 in the installation data table, and write the calculation result into the corresponding position in the installation data table;

[0030] Step 45a: Complete the comparison calculation of the last trackside device identified and the last trackside device in the installation data table; output the installation data table and end the step.

[0031] Furthermore, the processing of the scanned data in step 5 includes the following steps:

[0032] Step 51: Read the image data of the best image captured by the camera and the corresponding mileage data; and sort and label the acquired image data according to the mileage data;

[0033] Step 52: Compensate the mileage data corresponding to the image data of the best image to obtain compensated mileage data;

[0034] Step 53: Read a corresponding scan data based on the compensated mileage of the best image; when reading the scan data for the first time, the compensated mileage data of the first labeled image data is read; the scan data is read sequentially based on the compensated mileage data.

[0035] Step 54: Determine the position of the camera in the scanned image based on the height difference between the camera and the scanner, as well as the left and right deviations of the camera and the scanner in the direction perpendicular to the travel of the measuring vehicle;

[0036] Step 55: Calculate the camera's field of view in the scanned image;

[0037] Step 56: Extract the contour points of the trackside equipment's cross-section based on the camera's field of view within the scanned data;

[0038] Step 57: Based on the scanned contour points, fit the smallest bounding rectangle and calculate the center point of the bounding rectangle;

[0039] Step 58: Calculate the difference in x-coordinate between the center point of the circumscribed rectangle and the center point of the scanned image as the distance from the trackside device to the track in the scanned image; scale the distance in the scanned image according to the scanning ratio to obtain the actual distance and output it;

[0040] Step 59: Complete the processing of the scan data corresponding to the compensated mileage data of the last labeled image data, and end the step.

[0041] Furthermore, the compensation value in step 52 is the deviation distance between the scanner and the camera in the direction of travel of the measuring vehicle, and the compensation mileage = mileage - deviation distance.

[0042] Furthermore, in step 54, the height difference and left-right deviation between the camera and the scanner need to be converted to the scanned image by scanning ratio conversion; wherein the height difference ch in the scanned image = actual height difference * scanning ratio; the left-right deviation dh in the scanned image = actual left-right deviation * scanning ratio; the center point coordinates of the scanned image are represented as (Rx, Ry), and the position coordinates of the camera in the scanned image are represented as (Rx-dh, Ry-ch).

[0043] In step 55, L represents the scanning distance of the scanner, and H represents the maximum shooting range of the camera within the scanned image. Taking the camera's coordinate position (Rx-dh, Ry-ch) as vertex a, the target width of the camera as h, and the focal length of the lens as f, we obtain H = hL / f, and the coordinates of vertex b as (Rx-dh+L, Ry-ch+H / 2) and the coordinates of vertex c as (Rx-dh+L, Ry-ch-H / 2). A triangular region is drawn using vertices a, b, and c, and this triangular region represents the field of view of the camera in the scanned image.

[0044] The beneficial effects of this invention are as follows:

[0045] By setting up a camera to acquire images, and using a pre-trained model to identify the target equipment area in the selected image, the system determines whether the selected area is the best image based on its position in the image. For the best image, the corresponding mileage data is recorded as the actual mileage of the target equipment. After completing the mileage calculation for all trackside equipment, the clearance distance between adjacent trackside equipment is calculated to determine whether the engineering requirements are met.

[0046] By comparing the ID of the target device for optimal image recognition with a preset table in sequence, it can be determined whether there are any errors in the sequential installation of trackside equipment.

[0047] By comparing the mileage corresponding to the best image with a preset table in sequence, it can be determined whether the trackside equipment is installed at the preset mileage.

[0048] By compensating for mileage data, the scanner can be directly facing the target device at the compensated mileage data point, thus achieving accurate data acquisition.

[0049] By processing scan data that corresponds to the mileage data of the best image, the distance between the trackside equipment and the track can be obtained. Attached Figure Description

[0050] Figure 1 This is a schematic diagram of the structure of Embodiment 1 of the present invention;

[0051] Figure 2 This is a general flowchart of Embodiment 1 of the present invention;

[0052] Figure 3 This is a schematic diagram of the box selection for model training in Embodiment 1 of the present invention;

[0053] Figure 4 This is a schematic diagram of the process for calculating the clearance distance according to Embodiment 1 of the present invention;

[0054] Figure 5 This is a schematic diagram of the process for verifying the installation position of trackside equipment according to Embodiment 1 of the present invention;

[0055] Figure 6 This is a flowchart illustrating the distance calculation from the trackside equipment to the track according to Embodiment 1 of the present invention.

[0056] Figure 7 This is a schematic diagram of the camera position in a scanned image according to Embodiment 1 of the present invention;

[0057] Figure 8 This is a schematic diagram of the field of view of the camera in a scanned image according to Embodiment 1 of the present invention;

[0058] Figure 9 The distance is the horizontal axis distance from the center point of the selected area in the scanned image of Embodiment 1 of the present invention to the center point of the scanned image. Detailed Implementation

[0059] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0060] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0061] Example 1:

[0062] like Figure 1 As shown, an automatic trackside clearance measurement device includes a measuring vehicle, a laser scanner, a camera, a lens, an encoder, and a computing device. The laser scanner, camera, and encoder are all mounted on the measuring vehicle, and the lens is mounted on the camera. The laser scanner, camera, and encoder are all communicatively connected to the computing device. The measuring vehicle can be positioned on the track to be inspected and travels along the track's path.

[0063] The laser scanner is positioned in the middle of the measuring vehicle, with its scanning direction extending to both sides of the vehicle. In this example, the scanning direction is always perpendicular to the vehicle's direction of travel. The laser scanner is used to scan the trackside equipment and calculate the distance from the equipment to the track.

[0064] The camera is positioned at any point on the surveying vehicle. It acquires images along the vehicle's travel track. Subsequent software processing matches and calculates the image data with scanner data to verify the installation location of the trackside equipment and calculate the clearance distance, which is the interval distance. In this example, the camera's shooting direction is always perpendicular to the surveying vehicle's travel direction.

[0065] The encoder is installed on the wheels of the measuring vehicle and is used to record the travel distance of the measuring vehicle to assist in the positioning of the trackside equipment during detection.

[0066] The computing device can be a service terminal used to store and compute information data collected by cameras, laser scanners, and encoders.

[0067] During the implementation process, the measuring vehicle is set up on the track and moves along the track. The relevant position information of the trackside equipment is obtained by laser scanner and camera, and the travel distance of the measuring vehicle is obtained by encoder. Then, the limit distance of the trackside equipment and the specific mileage position of the trackside equipment relative to the track are obtained by computing device, thereby completing the detection of the trackside equipment.

[0068] like Figure 2 As shown, a method for rapid automatic measurement of trackside equipment clearance includes the following steps:

[0069] Step 1: The measuring vehicle receives the start signal and moves along the track. At the same time, the camera collects image information on both sides of the track, the laser scanner collects scanning data, and the encoder collects encoding data. The collected image data, scanning data, and encoding data are transmitted to the computing device in real time.

[0070] Step 2: The computing device receives image data and reads the simultaneously received encoded data, converts the encoded data into mileage data, and saves the image data and mileage data together; The computing device receives scan data and reads the simultaneously received encoded data, converts the encoded data into mileage data, and saves the scan data and mileage data together.

[0071] Step 3: The measuring vehicle receives a stop signal and stops moving forward on the track; the computing device begins to process the saved image data, scan data, and mileage data;

[0072] Step 4: The computing device completes image detection based on the received image data and corresponding mileage data;

[0073] Step 5: After image detection is completed, the computing device processes the scan data based on the received scan data and the corresponding mileage data to obtain the distance data from the trackside equipment to the track.

[0074] Step 6: Output the information of the trackside equipment and end the process.

[0075] The start signal in step 1 and the stop signal in step 3 are both remote control signals.

[0076] like Figure 3 , 4 As shown, the image detection in step 4 includes two parts: one is the calculation of the clearance distance between trackside equipment, and the other is the verification of the installation position of the trackside equipment. The calculation of the clearance distance includes the following steps:

[0077] Step 41: The model trained based on deep learning receives the input image data, traverses the image data, reads each frame of the image in sequence, and identifies and selects the target device in each frame of the image.

[0078] Step 42: Read the selected area in each frame of the image, and determine whether it is the best image based on the distance between the center point of the selected area and the corresponding center point of the image; determine the mileage corresponding to the target device based on the mileage data corresponding to the best image.

[0079] Step 43: Obtain the mileage corresponding to all track equipment, sort the track equipment according to the mileage, and calculate the clearance distance between adjacent equipment in terms of mileage.

[0080] Step 44: Organize and output the limit distances, then end the process.

[0081] In step 41, the deep learning-based model uses the YOLOv3 Darknet53 basic network model. During model training, a batch of images including the target device needs to be acquired first, and bounding boxes are processed to select the target device within the images using the smallest possible bounding box. The coordinates of any diagonal vertex of the bounding box are input. During training, the network model automatically learns the relative relationship between the bounding box and the pixel distribution of the image within the bounding box. After training, the output also includes the bounding box containing all pixels of the identified target, including the coordinates of the two vertices on any diagonal of the bounding box. The model training includes the following steps:

[0082] Step 411: Input training sample images and annotations, including the coordinates of any diagonal vertex of the selected region;

[0083] Step 412: Normalize the images in the training set and scale them to integer multiples of 32;

[0084] Step 413: Set the width and height of the initial candidate boxes in the boundary regression module;

[0085] Step 414: Train the network model using the preprocessed training set images;

[0086] Step 415: Input the image to be detected into the trained network model to detect the targets in the image, and output the detection results of the small target category and location;

[0087] Step 416: Filter the detection results according to the preset category confidence threshold and overlap threshold to obtain the final detection result;

[0088] Step 417: Generate and export the prediction model. After exporting the model, it will be called for inference when detecting subsequent images. After the model inference, the detection results of the target category and location in the image will be output, and the step will end.

[0089] The selection of the best image in step 42 includes the following steps:

[0090] Step 421: Read the selected area in the image and obtain the coordinates of two vertices on any diagonal of the selected area, denoted as (Sx, Sy) and (Ex, Ey) respectively; when reading the selected area of ​​the image for the first time, the selected area of ​​the first frame image is read.

[0091] Step 422: Calculate the coordinates of the midpoint of the two vertices (Sx, Sy) and (Ex, Ey), and represent the coordinates of the midpoint as (Cx, Cy).

[0092] Step 423: Obtain the coordinates of the center point of the image, and based on the coordinates of the midpoint of the selected area and the coordinates of the center point of the image, obtain the horizontal coordinate spacing DistanceCW between them;

[0093] Step 424: Determine whether to traverse all images; if so, proceed to step 425; otherwise, return to step 421.

[0094] Step 425: Traverse all images, select the frame with the smallest horizontal coordinate spacing DistanceCW as the best image, and end the step.

[0095] In step 423, the horizontal coordinate spacing is expressed as DistanceCW=|Cx-Width / 2|, where Width represents the length of the image in the horizontal coordinate direction.

[0096] In step 425, since the camera's viewing axis is perpendicular to the direction of travel of the measuring vehicle in this example, and the measuring vehicle is moving along the track, the mileage data of the selected area of ​​the target device that is closest to the center point of the image is the mileage corresponding to the trackside device.

[0097] like Figure 5 As shown, the verification of the trackside equipment installation location includes the following steps:

[0098] Step 41a: Obtain the trackside equipment and corresponding mileage data identified during the calculation of clearance distance; obtain the preset trackside equipment installation data table;

[0099] Step 42a: Number the trackside equipment according to the mileage data, and read the ID of the trackside equipment and the corresponding mileage data L1 in sequence; read the ID of the trackside equipment and the corresponding mileage data L2 in the table according to the serial number of the installation data table; the ID of the trackside equipment includes signal lights and speed measuring instruments.

[0100] Step 43a: Determine whether the trackside device ID obtained by identification matches the trackside device ID in the corresponding order in the installation data table; if the IDs match, write "Yes" in the corresponding position in the installation data table; otherwise, write "No".

[0101] Step 44a: Calculate the absolute value of the difference between the mileage data L1 obtained from the identification calculation and the mileage data L2 in the installation data table, and write the calculation result into the corresponding position in the installation data table;

[0102] Step 45a: Complete the comparison calculation of the last trackside device identified and the last trackside device in the installation data table; output the installation data table and end the step.

[0103] like Figure 6 As shown, the data processing in step 5 includes the following steps:

[0104] Step 51: Read the image data of the best image captured by the camera and the corresponding mileage data; and sort and label the acquired image data according to the mileage data;

[0105] Step 52: Compensate the mileage data corresponding to the image data of the best image to obtain compensated mileage data;

[0106] Step 53: Read a corresponding scan data based on the compensated mileage of the best image; when reading the scan data for the first time, the compensated mileage data of the first labeled image data is read; the scan data is read sequentially based on the compensated mileage data.

[0107] Step 54: Determine the position of the camera in the scanned image based on the height difference between the camera and the scanner, as well as the left and right deviations of the camera and the scanner in the direction perpendicular to the travel of the measuring vehicle;

[0108] Step 55: Calculate the camera's field of view in the scanned image;

[0109] Step 56: Extract the contour points of the trackside equipment's cross-section based on the camera's field of view within the scanned data;

[0110] Step 57: Based on the scanned contour points, fit the smallest bounding rectangle and calculate the center point of the bounding rectangle;

[0111] Step 58: Calculate the difference in x-coordinate between the center point of the circumscribed rectangle and the center point of the scanned image as the distance from the trackside device to the track in the scanned image; scale the distance in the scanned image according to the scanning ratio to obtain the actual distance and output it;

[0112] Step 59: Complete the processing of the scan data corresponding to the compensated mileage data of the last labeled image data, and end the step.

[0113] In step 52, due to the deviation in the installation positions of the camera and scanner, the mileage data between them will be inaccurate. Therefore, it is necessary to compensate for the scanned image data to improve measurement accuracy. The direction of travel of the measuring vehicle is taken as positive, and the compensation value is the deviation distance between the installation positions of the scanner and camera in the direction of travel of the measuring vehicle. Compensated mileage = mileage - deviation distance. It should be noted that if the scanner is installed in front of the camera, the deviation distance is positive; if the scanner is installed behind the camera, the deviation distance is negative. By compensating for the deviation distance, it is ensured that the scanner is directly facing the target device when the mileage is being compensated.

[0114] like Figure 7 As shown, in step 54, the height difference and left-right deviation between the camera and scanner need to be converted to a scanning scale when transferred to the scanned image. The height difference ch within the scanned image = actual height difference * scanning scale. Similarly, the left-right deviation dh within the scanned image = actual left-right deviation * scanning scale. If the center point coordinates of the scanned image are represented as (Rx, Ry), then the camera's position coordinates within the scanned image are represented as (Rx-dh, Ry-ch). In this example, the left-right deviation between the camera and scanner is 0.

[0115] like Figure 8 As shown, in step 55, the coordinate position of the camera in the scanned image is determined based on step 54. Let L represent the scanning distance of the scanner, and H represent the maximum shooting range of the camera within the scanned image. Using the camera's coordinate position (Rx-dh, Ry-ch) as vertex a, the camera's target width as h, and the lens focal length as f, we obtain H = hL / f. Therefore, the coordinates of vertex b are (Rx-dh+L, Ry-ch+H / 2), and the coordinates of vertex c are (Rx-dh+L, Ry-ch-H / 2). A triangular region is drawn using vertices a, b, and c, and this triangular region represents the field of view of the camera in the scanned image.

[0116] In step 56, the intersection points of the field of view area obtained in step 55 and the scanning points in the scanned image are calculated, and the set of all intersection points is output as the cross-sectional contour point set of the target device.

[0117] like Figure 9 As shown, the distance X from the trackside equipment to the track in step 58 is... d The calculation of is shown in the following formula:

[0118]

[0119] Where x represents the x-coordinate of the center point of the circumscribed rectangle in step 57, and using the center point of the circumscribed rectangle as the position of the target device can eliminate the influence of different trackside devices having different areas in the scanned image.

[0120] Actual distance X td , represented as X td =X d / S, where S represents the scanner's scanning ratio.

[0121] During implementation, images are acquired by a camera, and a pre-trained model identifies the target equipment area within the selected image. The position of the selected area in the image is used to determine if it is the optimal image. For the optimal image, the corresponding mileage data is recorded as the actual mileage of the target equipment. After calculating the mileage of all trackside equipment, the clearance distance between adjacent trackside equipment is calculated to determine if engineering requirements are met. The IDs of the target equipment identified in the optimal image are compared sequentially with a pre-set table to determine if there are any errors in the order of trackside equipment installation. The mileage corresponding to the optimal image is compared sequentially with the pre-set table to determine if the trackside equipment is installed at the pre-set mileage. Mileage data is compensated to ensure that the scanner is directly facing the target equipment at the compensated mileage data points. By processing scan data with the same mileage data as the optimal image, the distance between the trackside equipment and the track is obtained.

[0122] The above description is merely a specific example of the present invention and does not constitute any limitation on the present invention. Obviously, those skilled in the art, after understanding the content and principles of the present invention, may make various modifications and changes in form and detail without departing from the principles and structure of the present invention; however, these modifications and changes based on the spirit of the present invention are still within the scope of protection of the claims of the present invention.

Claims

1. A method for rapid and automatic measurement of trackside equipment clearance, characterized in that, Includes the following steps: Step 1: The measuring vehicle receives the start signal and moves along the track. At the same time, the camera collects image information on both sides of the track, the laser scanner collects scanning data, and the encoder collects encoding data. The collected image data, scanning data, and encoding data are transmitted to the computing device in real time. Step 2: The computing device receives image data and reads the simultaneously received encoded data, converts the encoded data into mileage data, and saves the image data and mileage data together; The computing device receives scan data and reads the simultaneously received encoded data, converts the encoded data into mileage data, and saves the scan data and mileage data together. Step 3: The measuring vehicle receives a stop signal and stops moving forward on the track; the computing device begins to process the saved image data, scan data, and mileage data; The start signal in step 1 and the stop signal in step 3 are both remote control signals; Step 4: The computing device completes image detection based on the received image data and corresponding mileage data; Step 5: After image detection is completed, the computing device processes the scan data based on the received scan data and the corresponding mileage data to obtain the distance data from the trackside equipment to the track. The data processing in step 5 includes the following steps: Step 51: Read the image data of the best image captured by the camera and the corresponding mileage data; and sort and label the acquired image data according to the mileage data; Step 52: Compensate the mileage data corresponding to the image data of the best image to obtain compensated mileage data; Step 53: Read a corresponding scan data based on the compensated mileage of the best image; when reading the scan data for the first time, the compensated mileage data of the first labeled image data is read; the scan data is read sequentially based on the compensated mileage data. Step 54: Determine the position of the camera in the scanned image based on the height difference between the camera and the scanner, as well as the left and right deviations of the camera and the scanner in the direction perpendicular to the travel of the measuring vehicle; Step 55: Calculate the camera's field of view in the scanned image; Step 56: Extract the contour points of the trackside equipment's cross-section based on the camera's field of view within the scanned data; Step 57: Based on the scanned contour points, fit the smallest bounding rectangle and calculate the center point of the bounding rectangle; Step 58: Calculate the difference in x-coordinate between the center point of the circumscribed rectangle and the center point of the scanned image as the distance from the trackside device to the track in the scanned image; scale the distance in the scanned image according to the scanning ratio to obtain the actual distance and output it; Step 59: Complete the processing of the scan data corresponding to the compensated mileage data of the last labeled image data, and end the step; Step 6: Output the information of the trackside equipment and end the process.

2. The method for rapid automatic measurement of trackside equipment clearance according to claim 1, characterized in that, The image detection in step 4 includes the calculation of the clearance distance between trackside equipment. The calculation of the clearance distance includes the following steps: Step 41: The model trained based on deep learning receives the input image data, traverses the image data, reads each frame of the image in sequence, and identifies and selects the target device in each frame of the image. Step 42: Read the selected area in each frame of the image, and determine whether it is the best image based on the distance between the center point of the selected area and the corresponding center point of the image; determine the mileage corresponding to the target device based on the mileage data corresponding to the best image. Step 43: Obtain the mileage corresponding to all track equipment, sort the track equipment according to the mileage, and calculate the clearance distance between adjacent equipment in terms of mileage. Step 44: Organize the boundary distances and output them in a table to complete the process.

3. The method for rapid automatic measurement of trackside equipment clearance according to claim 2, characterized in that, In step 41, the deep learning-based model uses the YOLOv3 Darknet53 basic network model.

4. The method for rapid automatic measurement of trackside equipment clearance according to claim 2, characterized in that, The selection of the best image in step 42 includes the following steps: Step 421: Read the selected area in the image and obtain the coordinates of two vertices on any diagonal of the selected area, denoted as (Sx, Sy) and (Ex, Ey) respectively; when reading the selected area of ​​the image for the first time, the selected area of ​​the first frame image is read. Step 422: Calculate the coordinates of the midpoint of the two vertices (Sx, Sy) and (Ex, Ey), and represent the coordinates of the midpoint as (Cx, Cy). Step 423: Obtain the coordinates of the center point of the image, and based on the coordinates of the midpoint of the selected area and the coordinates of the center point of the image, obtain the horizontal coordinate spacing DistanceCW between them; Step 424: Determine whether to traverse all images; if so, proceed to step 425. Otherwise, return to step 421; Step 425: Traverse all images, select the frame with the smallest horizontal coordinate spacing DistanceCW as the best image, and end the step.

5. The method for rapid automatic measurement of trackside equipment clearance according to claim 2, characterized in that, The image detection in step 4 also includes the verification of the trackside equipment installation location, which includes the following steps: Step 41a: Obtain the trackside equipment and corresponding mileage data identified during the calculation of clearance distance; Obtain the preset trackside equipment installation data table; Step 42a: Number the trackside equipment according to the mileage data, and read the ID of the trackside equipment and the corresponding mileage data L1 in sequence; read the ID of the trackside equipment and the corresponding mileage data L2 in the table according to the serial number of the installation data table; the ID of the trackside equipment includes signal lights and speed measuring instruments. Step 43a: Determine whether the trackside device ID obtained by identification matches the trackside device ID in the corresponding order in the installation data table; if the IDs match, write "Yes" in the corresponding position in the installation data table; Otherwise, write "No"; Step 44a: Calculate the absolute value of the difference between the mileage data L1 obtained from the identification calculation and the mileage data L2 in the installation data table, and write the calculation result into the corresponding position in the installation data table; Step 45a: Complete the comparison calculation of the last trackside device identified and the last trackside device in the installation data table; output the installation data table and end the step.

6. The method for rapid automatic measurement of trackside equipment clearance according to claim 1, characterized in that, The compensation value in step 52 is the deviation distance between the installation positions of the scanner and the camera in the direction of travel of the measuring vehicle, and the compensation mileage = mileage - deviation distance.

7. The method for rapid automatic measurement of trackside equipment clearance according to claim 1, characterized in that, In step 54, the height difference and left-right deviation between the camera and the scanner need to be converted to the scanned image by scanning ratio conversion; where the height difference ch in the scanned image = actual height difference * scan ratio; the left-right deviation dh in the scanned image = actual left-right deviation * scan ratio; the center point coordinates of the scanned image are represented as (Rx, Ry), and the position coordinates of the camera in the scanned image are represented as (Rx-dh, Ry-ch). In step 55, L represents the scanning distance of the scanner, and H represents the maximum shooting range of the camera within the scanned image. Taking the camera's coordinate position (Rx-dh, Ry-ch) as vertex a, the target width of the camera as h, and the focal length of the lens as f, we obtain H = hL / f, and the coordinates of vertex b as (Rx-dh+L, Ry-ch+H / 2) and the coordinates of vertex c as (Rx-dh+L, Ry-ch-H / 2). A triangular region is drawn using vertices a, b, and c, and this triangular region represents the field of view of the camera in the scanned image.

Images