A subway track surface extraction method, system, device and medium based on laser radar point cloud

Abstract

The application discloses a subway track surface extraction method, system, equipment and medium based on a laser radar point cloud, and the extraction method comprises the following steps: ground laser radar scans a subway tunnel to obtain subway tunnel point cloud data; a cylindrical fitting method is used to fit the tunnel point cloud data, and the tunnel point cloud data is separated into tunnel lining point cloud and remaining part point cloud; height filtering is performed on the remaining part point cloud to extract subgrade area point cloud; two-dimensional principal component analysis is used to perform coordinate conversion on the subgrade area point cloud; the subgrade area point cloud is rasterized; a sliding window is constructed to extract subway tunnel track surface point cloud of each grid column; and different grid column track surface area points are merged to obtain a current tunnel track surface. The application effectively extracts the subway tunnel track surface from a local perspective, avoids the problems of low applicability and unstable precision of a global method, improves the automation degree, extraction rate and accuracy of extraction, and realizes accurate extraction of the subway tunnel point cloud track surface.

Description

Technical Field

[0001] This invention relates to the field of remote sensing technology, and in particular to a method, system, device and medium for extracting subway track surfaces based on lidar point clouds. Background Technology

[0002] As a vital component of urban transportation, the geometric features of subway tunnel tracks are crucial for stable train operation and passenger comfort. Traditional track surface inspection methods often rely on manual measurement or sensor equipment, which are not only time-consuming and labor-intensive but also have extremely low spatial resolution, failing to comprehensively depict the condition of the subway tracks. LiDAR (Light Detection and Ranging) technology offers a new approach to subway track surface extraction. LiDAR can acquire large amounts of 3D point cloud data in a short time, and its high precision and density accurately reflect the internal geometry of subway tunnels. By processing and analyzing LiDAR point cloud data, rapid and comprehensive extraction and reconstruction of subway tunnel track surfaces can be achieved, providing reliable technical support for subway operation and maintenance, and bringing more efficient and reliable solutions for subway operation management and track maintenance. Summary of the Invention

[0003] Purpose of the invention: The present invention aims to provide a highly automated, high-precision extraction method for effectively identifying and extracting point clouds of subway track surfaces based on lidar; another purpose of the present invention is to provide a subway track surface extraction system, equipment and medium based on lidar point clouds.

[0004] Technical solution: The subway track surface extraction method based on lidar point cloud as described in this invention includes the following steps:

[0005] (1) Ground-based lidar scans the subway tunnel to obtain point cloud data of the subway tunnel;

[0006] (2) The MSAC cylindrical fitting method was used to fit the tunnel point cloud data and separate it into the tunnel lining point cloud pt and the remaining point cloud pl.

[0007] (3) Perform height filtering on the remaining point cloud pl to extract the point cloud pb of the roadbed area;

[0008] (4) Use two-dimensional principal component analysis to perform coordinate transformation on the point cloud pb of the roadbed area;

[0009] (5) Rasterized point cloud pb of the roadbed area;

[0010] (6) Construct a sliding window and extract the point cloud of the subway tunnel track surface for each grid column;

[0011] (7) Merge the points in different grid columns of the track surface to obtain the current tunnel track surface.

[0012] Furthermore, step (2) is as follows:

[0013] For subway tunnel point cloud data F(x) i ,y i ,z i ) i＝1,2,…,n The MSAC cylindrical fitting method was used for fitting, and the fitting equation was:

[0014]

[0015] Where x, y, and z are the three-dimensional coordinates of the tunnel scene surface, n is the number of points in the collected point cloud, i represents the i-th point, and r is the radius of the cylinder base. x ,a y ,a z (x0, y0, z0) is the axis vector of the cylinder, and (x0, y0, z0) is a point on the axis of the cylinder;

[0016] Substitute the tunnel point cloud data into the fitted cylinder and calculate the distance d from each point to the cylinder surface. i By setting a threshold distance d from the point to the cylindrical surface c The tunnel point cloud data is divided into tunnel lining point cloud pt and other part point cloud pl.

[0017] Furthermore, the remaining point cloud pl is subjected to height filtering, with the height filtering parameter being the median z-coordinate in pl.

[0018] Furthermore, step (4) is as follows:

[0019] Project the point cloud pb of the roadbed area onto the XOY plane to construct the corresponding covariance matrix C.

[0020]

[0021] Where m is the number of points in the point cloud pb of the roadbed region. For the XOY barycenter coordinates of the point set Perform principal component analysis on matrix C in the XOY plane to determine eigenvalues ​​λ1 and λ2. Arrange them in descending order to obtain the corresponding eigenvectors v1 and v2.

[0022] The coordinate transformation formula is as follows:

[0023]

[0024] Where pb1 represents the point cloud coordinates of the roadbed area after coordinate transformation.

[0025] Furthermore, step (5) is as follows:

[0026] Calculate the maximum and minimum values ​​of the X and Y coordinates in pb1, respectively. min ,x max ,y min ,y max ;

[0027] Set the grid size rs, and calculate the number of grid cells along the X and Y axes:

[0028] n x =floor((x max -x min ) / rs)

[0029] n y =floor((y max -y min ) / rs)

[0030] Where, n x and n y These represent the number of grid cells along the X and Y axes of the rasterized roadbed point cloud, respectively; floor(·) indicates rounding up.

[0031] Furthermore, step (6) is as follows:

[0032] Extract the points raster_coor within the k-th column of the raster based on the raster boundary. k,i Where i = 1, 2, ..., n y n y The number of grid cells contained in the k-th column.

[0033] Calculate the maximum z-coordinate of each grid point to obtain the sequence of maximum z-coordinates. Where i = 1, 2, ..., n y ;

[0034] Determine the maximum value in the sequence of maximum z-coordinates and the second largest value The current grid cell is the k1th or k2th cell in the current column.

[0035] Search for the point with the maximum elevation in the k1 and k2 grid cells, and define the sliding window length l using this point as the seed point. fix Slide along the positive and negative Y-axis respectively; extract the point cloud of the subway tunnel track surface for each grid column.

[0036] Furthermore, let's define the forward sliding to the k-th window, where the maximum z-coordinate of a point within the current window is zm. km The maximum z-coordinate of a point within the window before sliding is zm. km-1 If zm km-1 -zm km <zthreshold If the sliding stops, the point within the window from 1 km is the track surface point.

[0037] The subway track surface extraction system based on lidar point clouds described in this invention includes:

[0038] The acquisition module is used to acquire point cloud data of the subway tunnel, which is obtained by scanning the subway tunnel with ground lidar.

[0039] The fitting module is used to fit tunnel point cloud data;

[0040] The segmentation module is used to segment the subway tunnel point cloud data into tunnel lining point cloud (pt) and the remaining point cloud (pl).

[0041] The first extraction module is used to extract the roadbed area point cloud pb from the remaining point cloud pl.

[0042] The coordinate transformation module is used to perform coordinate transformation on the point cloud pb of the roadbed area;

[0043] The rasterization module is used to rasterize the point cloud (pb) of the roadbed area;

[0044] The second extraction module is used to extract the point cloud of the subway tunnel track surface for each grid column;

[0045] The merge module is used to merge points in different grid columns of the track surface to obtain the current tunnel track surface.

[0046] The computer device of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above method.

[0047] The computer-readable storage medium of the present invention stores a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0048] Beneficial effects: Compared with the prior art, the significant advantages of this invention are: 1. This invention overcomes the disadvantages of traditional point monitoring and realizes comprehensive analysis of subway tunnel tracks; 2. This invention has a high degree of automation and makes full use of the technical advantages of airborne LiDAR. The point cloud data it acquires has high accuracy, ensuring accurate and effective extraction of track surface point clouds while guaranteeing accuracy; 3. The extraction method described in this invention improves the accuracy of basic data for subsequent subway tunnel deformation monitoring. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the process of the present invention;

[0050] Figure 2 This is a scatter plot of a subway tunnel point cloud obtained in an embodiment of the present invention;

[0051] Figure 3 This is a scatter plot of the remaining point cloud after the removal of the tunnel lining in the present invention.

[0052] Figure 4 This is a scatter plot of the roadbed point cloud obtained by implementing the present invention;

[0053] Figure 5 This is a scatter plot of the roadbed point cloud after coordinate transformation according to the present invention;

[0054] Figure 6 This is a scatter plot of point cloud data of the subway tunnel track surface extracted for the purpose of implementing this invention. Detailed Implementation

[0055] The invention will now be further described with reference to the accompanying drawings.

[0056] like Figure 1 As shown, the subway track surface extraction method based on lidar point clouds of the present invention includes the following steps:

[0057] (1) Ground-based lidar scans the subway tunnel to obtain point cloud data of the subway tunnel;

[0058] (2) The MSAC cylindrical fitting method was used to fit the tunnel point cloud data and separate it into the tunnel lining point cloud pt and the remaining point cloud pl.

[0059] For subway tunnel point cloud data F(x) i ,y i ,z i ) i＝1,2,…,n The MSAC cylindrical fitting method was used for fitting, and the fitting equation was:

[0060]

[0061] Where x, y, and z are the three-dimensional coordinates of the tunnel scene surface, n is the number of points in the collected point cloud, i represents the i-th point, and r is the radius of the cylinder base. x ,a y ,a z (x0, y0, z0) is the axis vector of the cylinder, and (x0, y0, z0) is a point on the axis of the cylinder;

[0062] Substitute the tunnel point cloud data into the fitted cylinder and calculate the distance d from each point to the cylinder surface. i By setting a threshold distance d from the point to the cylindrical surface c The tunnel point cloud data is divided into tunnel lining point cloud pt and other part point cloud pl.

[0063] (3) Perform height filtering on the remaining point cloud pl to extract the roadbed area point cloud pb; perform height filtering on the remaining point cloud pl, with the height filtering parameter being the median value of the z coordinate in pl.

[0064] (4) Use two-dimensional principal component analysis to perform coordinate transformation on the point cloud pb of the roadbed area;

[0065] Project the point cloud pb of the roadbed area onto the XOY plane to construct the corresponding covariance matrix C.

[0066]

[0067] Where m is the number of points in the point cloud pb of the roadbed region. For the XOY barycenter coordinates of the point set Perform principal component analysis on matrix C in the XOY plane to determine eigenvalues ​​λ1 and λ2. Arrange them in descending order to obtain the corresponding eigenvectors v1 and v2.

[0068] The coordinate transformation formula is as follows:

[0069]

[0070] Where pb1 represents the point cloud coordinates of the roadbed area after coordinate transformation.

[0071] (5) Rasterized point cloud pb of the roadbed area;

[0072] Calculate the maximum and minimum values ​​of the X and Y coordinates in pb1, respectively. min ,x max ,y min ,y max ;

[0073] Set the grid size rs, and calculate the number of grid cells along the X and Y axes:

[0074] n x =floor((x max -x min ) / rs)

[0075] n y =floor((y max -y min ) / rs)

[0076] Where, n x and n y These represent the number of grid cells along the X and Y axes of the rasterized roadbed point cloud, respectively; floor(·) indicates rounding up.

[0077] (6) Construct a sliding window and extract the point cloud of the subway tunnel track surface for each grid column;

[0078] Extract the points raster_coor within the k-th column of the raster based on the raster boundary. k,i Where i = 1, 2, ..., n y n y The number of grid cells contained in the k-th column.

[0079] Calculate the maximum z-coordinate of each grid point to obtain the sequence of maximum z-coordinates. Where i = 1, 2, ..., n y ;

[0080] Determine the maximum value in the sequence of maximum z-coordinates and the second largest value The current grid cell is the k1th or k2th cell in the current column.

[0081] Search for the point with the maximum elevation in the k1 and k2 grid cells, and define the sliding window length l using this point as the seed point. fix Slide along the positive and negative Y-axis respectively; extract the point cloud of the subway tunnel track surface for each grid column.

[0082] Assuming the forward sliding reaches the k-th window, the maximum z-coordinate of a point within the current window is zm. km The maximum z-coordinate of a point within the window before sliding is zm. km-1 If zm km-1 -zm km <z threshold If the sliding stops, the point within the window from 1 km is the track surface point.

[0083] (7) Merge the points in different grid columns of the track surface to obtain the current tunnel track surface.

[0084] The subway track surface extraction system based on lidar point clouds described in this invention includes:

[0085] The acquisition module is used to acquire point cloud data of the subway tunnel, which is obtained by scanning the subway tunnel with ground lidar.

[0086] The fitting module is used to fit tunnel point cloud data;

[0087] The segmentation module is used to segment the subway tunnel point cloud data into tunnel lining point cloud (pt) and the remaining point cloud (pl).

[0088] The first extraction module is used to extract the roadbed area point cloud pb from the remaining point cloud pl.

[0089] The coordinate transformation module is used to perform coordinate transformation on the point cloud pb of the roadbed area;

[0090] The rasterization module is used to rasterize the point cloud (pb) of the roadbed area;

[0091] The second extraction module is used to extract the point cloud of the subway tunnel track surface for each grid column;

[0092] The merge module is used to merge points in different grid columns of the track surface to obtain the current tunnel track surface.

[0093] The computer device of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above method.

[0094] The computer-readable storage medium of the present invention stores a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0095] Taking the point cloud data obtained from a roadside in Zhenjiang, Jiangsu Province, using the Hummingbird Genius micro-drone equipped with a LiDAR system as an example, the specific operation steps are as follows:

[0096] (1) Use ground-based lidar to scan the subway tunnel and obtain tunnel point cloud data F(x) i ,y i ,z i ) i＝1,2,…,n Where x, y, and z are the three-dimensional coordinates of the tunnel scene surface, n is the number of points in the collected point cloud, and i represents the i-th point; the scatter plot within the scene is as follows: Figure 2 As shown;

[0097] (2) Point cloud F(x) of the subway tunnel scene i ,y i ,z i ) i＝1,2,…,n Using the MSAC cylindrical fitting method, the fitting model can be expressed as:

[0098]

[0099] In the formula, r is the radius of the cylinder's base, (a x ,a y ,a z (x0, y0, z0) is the axis vector of the cylinder, and (x0, y0, z0) is a point on the axis of the cylinder.

[0100] Substitute each point into the fitted cylindrical method to calculate the distance d from each point to the cylindrical surface. i Set the distance threshold d from the point to the model. c Separate the point cloud pt of the tunnel lining and the point cloud pl of the remaining parts, such as Figure 3 As shown.

[0101] (3) Perform height filtering on the remaining point cloud pl, generally using the median z-coordinate in pl as the height filtering parameter, to obtain the roadbed area point cloud pb, such as... Figure 4 As shown.

[0102] (4) Perform two-dimensional principal component analysis on the point cloud of the roadbed area to transform the coordinates, and obtain the coordinate transformation matrix as follows: The point cloud scatter plot after coordinate transformation is as follows Figure 5 As shown.

[0103] (5) Set the grid size to 0.3m, and the number of grids in the X-axis and Y-axis directions to 69 and 11 respectively.

[0104] (6) Construct a sliding window to extract the subway tunnel track surface. Taking the third column as an example, the third column contains 11 grids. Extract the points in each grid according to the grid boundary.

[0105] Calculate the maximum z-coordinate of each of the above grid points to obtain the sequence of maximum z-coordinates: Where i = 1, 2, ..., 11.

[0106] The grid cells containing the maximum and second largest Z-coordinate values ​​in the maximum Z-coordinate sequence were calculated. Their corresponding maximum Z-coordinate values ​​were -1.2098m and -1.2156m, respectively, and the corresponding grid cells were the 4th and 9th in the current column.

[0107] Search for the point with the maximum elevation in the 4th and 9th grid cells, and define the sliding window length l using this point as the seed point. fix =0.05, slide along the positive and negative Y-axis respectively, and set the maximum difference of the Z coordinate of adjacent sliding windows to 0.01m to obtain the point cloud of the current grid column orbit surface.

[0108] (7) By merging the point clouds of the track surfaces from all grid columns, the current subway tunnel track surface point cloud can be obtained, such as... Figure 6 As shown.

Claims

1. A method for extracting subway track surfaces based on lidar point clouds, characterized in that, Includes the following steps: (1) Ground-based lidar scans the subway tunnel to obtain point cloud data of the subway tunnel; (2) The tunnel point cloud data was fitted using the MSAC cylindrical fitting method. The tunnel point cloud data was substituted into the fitted cylinder, and the distance from each point to the cylinder surface was calculated. By setting a threshold distance from a point to the cylindrical surface The tunnel point cloud data is divided into tunnel lining point clouds. and the remaining point clouds ; (3) For the remaining point cloud Perform height filtering, with the height filtering parameters as follows: Extracting the point cloud of the roadbed area from the median z-coordinate. ; (4) Using two-dimensional principal component analysis to analyze the point cloud of the roadbed area Perform coordinate transformation; (5) Rasterized roadbed area point cloud ; (6) Construct a sliding window and extract the point cloud of the subway tunnel track surface for each grid column; (7) Merge the points in different grid columns of the track surface to obtain the current tunnel track surface; Step (4) is as follows: Point cloud of the roadbed area Projecting onto the XOY plane, construct the corresponding covariance matrix C. Where m is the point cloud of the roadbed region. The number of points in , For the XOY barycenter coordinates of the point set For the matrix Perform principal component analysis in the XOY plane to determine eigenvalues. and Arranged in descending order, the corresponding feature vectors are as follows: and ; The coordinate transformation formula is as follows: in, These are the point cloud coordinates of the roadbed area after coordinate transformation.

2. The method for extracting subway track surfaces based on lidar point clouds according to claim 1, characterized in that, Step (2) is as follows: Point cloud data of subway tunnels The MSAC cylindrical fitting method was used for fitting, and the fitting equation was: Where x, y, and z are the three-dimensional coordinates of the tunnel scene surface. The number of points collected in the point cloud, where i represents the i-th point. Let be the radius of the cylinder's base. The axis vector of the cylinder. Let be a point on the axis of the cylinder.

3. The method for extracting subway track surface based on lidar point clouds according to claim 1, characterized in that, Step (5) is as follows: calculate The maximum and minimum values ​​of the X and Y coordinates are respectively , , , ; Set grid size Calculate the number of grid cells along the X and Y axes: in, and These represent the number of grid cells along the X and Y axes of the rasterized roadbed point cloud, respectively. This indicates rounding up to the nearest integer.

4. The method for extracting subway track surface based on lidar point clouds according to claim 3, characterized in that, Step (6) is as follows: Extracting the first from the grid boundary Points within a column grid ,in, , For the first The number of cells contained in a column grid; Calculate each grid point The maximum value of the coordinates is obtained. Maximum coordinate sequence ,in, ; Sure Maximum value in the sequence of maximum coordinates and the second largest value The current grid cell, whose position is the first cell in the current column. and indivual; Search the and The point with the maximum elevation in the raster is used as the seed point to define the length of the sliding window. Slide along the positive and negative Y-axis respectively; extract the point cloud of the subway tunnel track surface for each grid column.

5. The method for extracting subway track surface based on lidar point clouds according to claim 4, characterized in that, Set forward slide to the first A window, the point within the current window The maximum value of the coordinates is Before sliding, the point inside the window The maximum value of the coordinates is ,like Then the sliding stops, and window 1 to... The interior point is the point on the orbital surface.

6. A subway track surface extraction system based on the method for extracting subway track surfaces using lidar point clouds according to any one of claims 1-5, characterized in that, include The acquisition module is used to acquire point cloud data of the subway tunnel, which is obtained by scanning the subway tunnel with ground lidar. The fitting module is used to fit tunnel point cloud data; The segmentation module is used to segment subway tunnel point cloud data into tunnel lining point clouds. and the remaining point clouds ; The first extraction module is used to extract data from the remaining point cloud. Extracting point clouds of roadbed areas ; The coordinate transformation module is used to transform the point cloud of the roadbed area. Perform coordinate transformation; Rasterization module, used for rasterizing point clouds of roadbed areas. ; The second extraction module is used to extract the point cloud of the subway tunnel track surface for each grid column; The merge module is used to merge points in different grid columns of the track surface to obtain the current tunnel track surface.

7. 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 computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images