Tunnel bending area point cloud acquisition device and coordinate correction method
By combining a two-dimensional cross-sectional laser scanner, an incremental ranging encoder, and a surface array lidar with the SLAM algorithm, the three-dimensional point cloud data of the tunnel's curved area is corrected, solving the positional error problem caused by the offset of the scanning device on the inspection vehicle and improving the accuracy of tunnel defect detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA UNIV OF MINING & TECH (BEIJING)
- Filing Date
- 2024-10-16
- Publication Date
- 2026-04-17
AI Technical Summary
In the curved areas of subway tunnels, the positional error caused by the offset between the detection vehicle's laser radar scanning device and the tunnel center affects the accuracy of the 3D laser scanning point cloud, making it difficult to achieve precise detection.
By employing a two-dimensional cross-sectional laser scanner, an incremental ranging encoder, an area array lidar, and a mobile terminal, combined with the SLAM algorithm, the three-dimensional point cloud data is corrected and positional errors are eliminated through centerline extraction and B-spline interpolation fitting.
It improved the detection accuracy of defects in curved areas of tunnels, and enabled precise positioning of target objects and construction of three-dimensional models within the tunnel.
Smart Images
Figure CN121878718A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of tunnel defect detection equipment technology, and more specifically, to a point cloud acquisition device and coordinate correction method for curved areas of tunnels. Background Technology
[0002] Currently, the detection of safety hazards in subway tunnels in my country has entered the stage of vehicle-mounted integrated intelligent detection. A key aspect of this comprehensive detection is acquiring tunnel point cloud data using 3D laser scanning technology to identify potential tunnel defects. However, the enclosed and narrow spatial environment of subway tunnels makes it impossible to use GNSS (Global Navigation Satellite System) signals for positioning correction within the tunnel. Furthermore, the curvature of curved areas in tunnels is mostly smooth, making it difficult to correct simply by adding curvature information factors. Therefore, eliminating the positional error caused by the offset between the scanning settings of the tunnel's laser array and the tunnel center when the tunnel inspection vehicle passes through curved areas, and improving the accuracy of the point cloud acquired by 3D laser scanning, is a crucial prerequisite for achieving accurate detection of safety hazards in subway tunnels and a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0003] The technology proposed in this invention is a point cloud acquisition device and coordinate correction method for curved areas of tunnels. The purpose is to eliminate the positional error in point cloud acquisition data caused by the offset between the tunnel inspection vehicle's laser radar scanning device and the tunnel center, thereby improving the detection accuracy of defects in curved areas of tunnels.
[0004] This invention provides a point cloud acquisition device for curved areas of tunnels, characterized in that it includes a two-dimensional cross-section laser scanner, an incremental ranging encoder, a surface array lidar, and a mobile terminal, and uses a tunnel inspection vehicle as a mobile platform to acquire two-dimensional cross-section point cloud, inspection vehicle mileage, and three-dimensional point cloud data information of the tunnel.
[0005] The two-dimensional cross-section laser scanner is installed at the rear of the tunnel inspection vehicle. In the working state, in order to ensure that the scanner's field of view is unobstructed, a connecting rod extending from the rear of the inspection vehicle is set to fix it, and high-density tunnel cross-section point cloud information is obtained by scanning.
[0006] The incremental ranging encoder is installed on the wheels of the tunnel inspection vehicle;
[0007] The area array lidar and mobile terminal are located inside the tunnel inspection vehicle. During operation, the scanning angle of the area array lidar is flexibly controlled by hand to obtain comprehensive point cloud model data, and the data acquisition is controlled by wired connection through the mobile terminal.
[0008] Furthermore, the aforementioned point cloud acquisition device for curved areas of tunnels is characterized in that, during tunnel inspection, the two-dimensional cross-section laser scanner is controlled by a mobile terminal to acquire two-dimensional cross-sectional point cloud information of the tunnel.
[0009] Furthermore, the aforementioned point cloud acquisition device for curved tunnel areas is characterized by an incremental ranging encoder capable of measuring the distance traveled by a tunnel inspection vehicle within the tunnel; these mileage values provide third-dimensional positional information. During data acquisition, the encoder synchronization signal is output to the scanner via a data cable, synchronizing the mileage values provided by the encoder with the two-dimensional cross-sectional point cloud data acquired by the scanner in real time. Through this synchronization, the two-dimensional cross-sectional point cloud coordinates can be transformed into complete three-dimensional coordinates, accurately reflecting the spatial position of the target object within the tunnel.
[0010] Furthermore, the aforementioned point cloud acquisition device for curved tunnel areas is characterized in that the area array lidar device is connected to a mobile terminal via a network cable, and the SLAM (Simultaneous Localization and Mapping) algorithm in the mobile terminal is used to control the radar data acquisition process. Subsequently, the SLAM algorithm is used to construct a map, forming a relevant three-dimensional point cloud model. This integrated design enables the area array lidar to achieve data acquisition, processing, and three-dimensional model construction with the support of a mobile terminal.
[0011] This invention provides a method for correcting point cloud coordinates in a curved area of a tunnel, characterized by the following steps: ① Tunnel point cloud data acquisition: using a tunnel inspection vehicle as a mobile platform, tunnel point cloud data is acquired through a two-dimensional cross-section scanner and a surface-array lidar; ② Obtaining curvature parameters of the curved area of the tunnel: obtaining spatial point cloud coordinate data of the tunnel through a centerline extraction algorithm, and fitting it to the tunnel's central axis to quickly identify the curved area of the tunnel and its curvature radius parameters; ③ Centerline trajectory correction: using a point cloud difference mapping algorithm, the tunnel curvature parameters obtained by SLAM (Simultaneous Localization and Mapping) are used to correct the point cloud axis information obtained by three-dimensional laser scanning; ④ Point cloud data correction: the two-dimensional cross-section point cloud data is corrected through a point cloud correction algorithm to obtain accurate cross-sectional point cloud coordinates of the curved area of the tunnel.
[0012] Furthermore, the method for correcting point cloud data in the curved area of the tunnel inspection vehicle is characterized in that: in step ①, two sets of point cloud data for the curved area of the tunnel are acquired using a two-dimensional cross-section scanner and a surface-array LiDAR. Three-dimensional laser scanning technology is achieved through a cross-section scanner installed at the rear of the tunnel inspection vehicle. During tunnel inspection, the scanner drives the laser emitter to rotate 360° via a rotating shaft. Since the rotation speed of the three-dimensional laser scanner reaches 200 revolutions per second, and the operating speed of the inspection vehicle is much lower than the rotation speed, three-dimensional point cloud model data is obtained by superimposing mileage, but it lacks expression of curvature information for the curved area of the tunnel. SLAM laser scanning technology uses a surface-array LiDAR connected to a power source and a mobile PC to construct the tunnel point cloud map in real time, and uses a handheld radar device while seated on the inspection vehicle to collect point cloud data.
[0013] Furthermore, the method for correcting point cloud data in the curved area of the tunnel inspection vehicle is characterized in that: in step ②, the curvature information extraction of the curved area of the tunnel firstly involves processing the point cloud data obtained by SLAM laser scanning technology, combining a two-dimensional projection strategy and angle criteria to detect tunnel boundary points, estimating two boundary lines on the XY plane, and then obtaining the tunnel's spatial central axis through the two boundary lines. The mathematical expression y = f(x) of the obtained axis is analyzed, and further, the curvature formula is used: Calculate the curvature information at each point, and use the relationship of curvature changes to determine the location of the tunnel's curved areas.
[0014] Furthermore, the method for correcting point cloud data in the curved area of the tunnel inspection vehicle is characterized in that: in step ③, the coordinates of the central axis of the point cloud obtained by the three-dimensional laser scanning are corrected by interpolation mapping using the curvature parameters of the curved tunnel. Specifically, uniform control points on the central axis are selected for mapping correction, and then a cubic B-spline interpolation fitting algorithm is used to fit the control points to obtain the corrected three-dimensional laser scanning central axis. B-spline fitting is a curve fitting method based on Bézier curves. It divides the data points into multiple intervals and fits a low-order polynomial or B-spline function within each interval to approximate the actual curve. Its advantages include high smoothness and approximation accuracy, while effectively reducing oscillations caused by global interpolation. The order of the B-spline fitting is determined based on the data characteristics and the required smoothness of the curve; in this correction, a third-order curve is selected to meet the accuracy requirements.
[0015] Furthermore, the method for correcting point cloud data in the curved area of the tunnel inspection vehicle is characterized by the following: in step ④, point cloud data correction involves analyzing and establishing the spatial geometric coordinate relationship between the central axis point cloud trajectory and the original two-dimensional cross-sectional point cloud, and correcting the cross-sectional point cloud coordinates of the curved area of the tunnel based on the curvature information obtained in step ②. Afterwards, the corrected two-dimensional cross-sections are combined to form a three-dimensional point cloud model of the curved area of the tunnel. Attached Figure Description
[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0017] Figure 1 A flowchart for the point cloud coordinate correction method in the curved area of a tunnel.
[0018] Figure 2 This is a schematic diagram illustrating the source of positional error in this invention.
[0019] Figures (a) and (b) provide a detailed explanation of the positional error. Figure (a) is a simplified diagram of the position of the inspection vehicle when traveling in the curved area of the tunnel. The yellow square represents the vehicle body, the dashed line represents the tunnel centerline, L is the distance between the scanner and the center of the rear wheel of the inspection vehicle, and the blue square at the rear is the cross-section scanner. To ensure a wide scanning field of view, it is mounted on the telescopic rod at the rear of the inspection vehicle. When the inspection vehicle travels in the curved area of the tunnel, the scanner will deviate from the tunnel centerline, further causing a positional error between the tunnel contour point cloud section CD acquired by the scanner and the actual tunnel contour interface AB. θ is the angular deviation between section AB and section CD.
[0020] Figure (b) is a schematic diagram for calculating the position error. According to the geometric relationship θ = arctan(L / R), the position error d can be expressed as: If the tunnel turning radius R is known to be 2000m and the distance L between the 2D cross-section laser scanner and the center of the vehicle is 1m, then the angle deviation can be calculated as follows: Position error Since the deformation of tunnel structures is typically on the order of millimeters, this positional error cannot be ignored.
[0021] Figure 3 This is a diagram showing the mapping relationship of the central axis.
[0022] Among them, such as Figure 3 As shown, steps ① and ② identify the curved areas of the tunnel and acquire the starting point cloud data of these curved areas. Uniform control points are selected from the central axis obtained by 3D laser scanning, and the curvature radius information of each control point is obtained through the interpolation mapping relationship illustrated in the diagram. Thus, the corrected central axis curve trajectory is obtained through curve fitting. Detailed Implementation
[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0024] The purpose of this invention is to provide a point cloud acquisition device and coordinate correction method for curved areas of tunnels, in order to eliminate the positional error in point cloud acquisition data caused by the offset between the tunnel inspection vehicle's laser radar scanning device and the tunnel center, thereby improving the detection accuracy of defects in curved areas of tunnels.
[0025] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0026] Please refer to Figure 1 , 2 and 3, Figure 1 Flowchart of the point cloud coordinate correction method for curved areas of tunnels. Figure 2 This is a schematic diagram illustrating the sources of positional error. Figure 3 This is a diagram showing the mapping relationship of the central axis.
[0027] like Figure 2 As shown in (a), the present invention provides a method for correcting point cloud coordinates in a curved area of a tunnel, characterized by the following steps: ① Tunnel point cloud data acquisition: using a tunnel inspection vehicle as a mobile platform, tunnel point cloud data is acquired through a two-dimensional cross-section scanner and a surface-array lidar; ② Obtaining curvature parameters of the curved area of the tunnel: obtaining SLAM (Simultaneous Localization and Mapping) tunnel spatial point cloud coordinate data through a centerline extraction algorithm, and fitting it to the tunnel centerline to quickly identify the curved area of the tunnel and its curvature radius parameters; ③ Centerline trajectory correction: using a point cloud difference mapping algorithm, the tunnel curvature parameters obtained by SLAM are used to correct the point cloud axis information obtained by three-dimensional laser scanning; ④ Point cloud data correction: the two-dimensional cross-section point cloud data is corrected through a point cloud correction algorithm to obtain accurate cross-sectional point cloud coordinates of the curved area of the tunnel.
[0028] In step ①, two sets of point cloud data for the curved areas of the tunnel are acquired using a two-dimensional cross-section scanner and a surface-array LiDAR. Three-dimensional laser scanning technology is implemented using a cross-section scanner installed at the rear of the tunnel inspection vehicle, with data acquisition, control, and storage handled in the vehicle's control room. SLAM laser scanning technology uses a surface-array LiDAR connected to a power source and a mobile PC to construct the tunnel point cloud map in real time, with point cloud data acquisition and storage performed by a handheld radar device while seated on the inspection vehicle.
[0029] In step ②, the curvature information extraction of the tunnel's curved area involves several steps. First, the point cloud data acquired using SLAM laser scanning technology is processed. Tunnel boundary points are detected using a two-dimensional projection strategy and angle criteria. Curve fitting is then performed on the boundary point cloud data to obtain two boundary lines on the XY plane. Next, the tunnel's spatial central axis is obtained using these two boundary lines. The mathematical expression for the obtained axis, y = f(x), is analyzed, and further processed using the curvature formula: Calculate the curvature information at each point, and use the relationship of curvature changes to determine the location of the tunnel's curved areas.
[0030] like Figure 3 As shown, in step ③, the coordinates of the central axis obtained by the three-dimensional laser scanning are corrected by interpolation mapping using the curvature parameters of the curved tunnel. Specifically, if the distance of the curved area of the tunnel is 1000m, the control point density is 10 control points within 1m, that is, 10001 control points are selected at a distance of 1000m. The point cloud coordinate information is obtained by SLAM to map and correct the control points. Then, the cubic B-spline interpolation fitting algorithm is used to perform curve fitting on the control points to obtain the corrected three-dimensional laser scanning central axis.
[0031] In step ④, point cloud data correction involves analyzing and establishing the spatial geometric coordinate relationship between the central axis point cloud trajectory and the original two-dimensional cross-sectional point cloud, and correcting the cross-sectional point cloud coordinates of the tunnel curvature area based on the curvature information obtained in step ②. Then, the corrected two-dimensional cross-sections are combined to form a three-dimensional point cloud model of the tunnel curvature area, thereby enabling the detection of defects in the tunnel curvature area.
[0032] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although modifications may be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions may be made to some of the technical features, any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A point cloud acquisition device for curved areas of tunnels, characterized in that: It includes a two-dimensional cross-section laser scanner, an incremental ranging encoder, a surface array lidar, and a mobile terminal. It uses a tunnel inspection vehicle as a mobile platform to acquire two-dimensional cross-section point cloud, inspection vehicle mileage, and tunnel three-dimensional point cloud data information. The two-dimensional cross-section laser scanner is installed at the rear of the tunnel inspection vehicle. In the working state, in order to ensure that the scanner's field of view is unobstructed, a connecting rod extending from the rear of the inspection vehicle is set to fix it, and high-density tunnel cross-section point cloud information is obtained by scanning. The incremental ranging encoder is installed on the wheels of the tunnel inspection vehicle; The area array lidar and mobile terminal are located inside the tunnel inspection vehicle. During operation, the scanning angle of the area array lidar is controlled by a handheld device to obtain comprehensive point cloud model data, and the data acquisition is controlled by a wired connection through the mobile terminal.
2. The point cloud acquisition device for a curved area in a tunnel according to claim 1, characterized in that, During tunnel inspection, the aforementioned two-dimensional cross-section laser scanner is controlled by a mobile terminal to acquire two-dimensional cross-sectional point cloud information of the tunnel.
3. The point cloud acquisition device for a curved area in a tunnel according to claim 1, characterized in that, The incremental ranging encoder described above can measure the distance traveled by a tunnel inspection vehicle within a tunnel, providing third-dimensional positional information. During data acquisition, the encoder synchronization signal is output to the scanner via a data cable, synchronizing the encoder's mileage values with the two-dimensional cross-sectional point cloud data acquired by the scanner in real time. This synchronization transforms the two-dimensional cross-sectional point cloud coordinates into complete three-dimensional coordinates, accurately reflecting the spatial position of target objects within the tunnel.
4. The point cloud acquisition device for a curved area in a tunnel according to claim 1, characterized in that, The described area-array lidar device is connected to a mobile terminal via a network cable and utilizes the SLAM (Simultaneous Localization and Mapping) algorithm in the mobile terminal to control the radar data acquisition process. Subsequently, the SLAM algorithm is used to construct a corresponding 3D point cloud model. This integrated design enables the area-array lidar to achieve data acquisition, processing, and 3D model construction with the support of a mobile terminal.
5. A method for correcting point cloud coordinates in a curved area of a tunnel, characterized in that: Includes the following steps: ① Tunnel point cloud data acquisition: Using a tunnel inspection vehicle as a mobile platform, tunnel point cloud data is acquired through a two-dimensional cross-section scanner and a surface array lidar. ② Obtaining curvature parameters of the tunnel's curved areas: The spatial point cloud coordinates of the tunnel are obtained through a centerline extraction algorithm and fitted to the tunnel's centerline to quickly identify the curved areas of the tunnel and obtain their curvature radius parameters; ③ Centerline trajectory correction: The tunnel curvature parameters obtained by SLAM are used to correct the point cloud centerline information obtained by 3D laser scanning technology through a point cloud difference mapping algorithm; ④ Point cloud data correction: The 2D cross-sectional point cloud data obtained by 3D laser scanning technology is corrected a second time through a point cloud correction algorithm to obtain the accurate cross-sectional point cloud coordinates of the tunnel's curved areas.
6. The method for correcting point cloud coordinates in a curved area of a tunnel according to claim 5, characterized in that: This invention can eliminate the positional error in point cloud data acquisition caused by the offset between the tunnel inspection vehicle's array lidar scanning device and the tunnel center, thereby improving the detection accuracy of defects in curved areas of the tunnel.
7. The method for correcting point cloud coordinates in a curved area of a tunnel according to claim 5, characterized in that: In step ①, two sets of point cloud data for the curved areas of the tunnel are acquired using a two-dimensional cross-section scanner and a surface-array LiDAR. Three-dimensional laser scanning technology is implemented using a two-dimensional cross-section scanner installed at the rear of the tunnel inspection vehicle. During tunnel inspection, the scanner drives the laser emitter to rotate 360° via a rotating shaft. Since the three-dimensional laser scanner rotates at 200 revolutions per second, and the inspection vehicle's speed is much lower than this speed, three-dimensional point cloud model data is obtained by overlaying mileage data. However, this data lacks information on the curvature of the curved areas of the tunnel. SLAM laser scanning technology uses a surface-array LiDAR connected to a power source and a mobile PC to construct the tunnel point cloud map in real time. Point cloud data is collected using a handheld radar device while seated on the inspection vehicle.
8. The method for correcting point cloud coordinates in a curved area of a tunnel according to claim 5, characterized in that: In step ②, the curvature information extraction for the tunnel's curved area involves several steps. First, the point cloud data acquired using SLAM laser scanning technology is processed. Then, tunnel boundary points are detected using a two-dimensional projection strategy and angle criteria to estimate two boundary lines on the XY plane. Next, the tunnel's spatial central axis is obtained using these two boundary lines. The mathematical expression for the obtained axis, y = f(x), is analyzed, and further processed using the curvature formula: Calculate the curvature information at each point, and use the relationship of curvature changes to determine the location of the tunnel's curved areas.
9. The method for correcting point cloud coordinates in a curved area of a tunnel according to claim 5, characterized in that: In step ③, the coordinates of the central axis obtained from the 3D laser scanning are corrected by interpolation mapping using the curvature parameters of the curved tunnel. Specifically, uniform control points on the central axis are selected for mapping correction, and then a cubic B-spline interpolation fitting algorithm is used to fit the control points to obtain the corrected 3D laser scanning central axis. B-spline fitting is a curve fitting method based on Bézier curves. It divides the data points into multiple intervals and fits a low-order polynomial or B-spline function within each interval to approximate the actual curve. Its advantages include high smoothness and approximation accuracy, while effectively reducing oscillations caused by global interpolation. The order of the B-spline fitting is determined based on the data characteristics and the required smoothness of the curve; in this correction, a third-order curve is selected to meet the accuracy requirements.
10. The method for correcting point cloud coordinates in a curved area of a tunnel according to claim 5, characterized in that: In step ④, point cloud data correction involves analyzing and establishing the spatial geometric coordinate relationship between the central axis point cloud trajectory and the original two-dimensional cross-sectional point cloud, and correcting the cross-sectional point cloud coordinates of the tunnel curvature area based on the curvature information obtained in step ②. Then, the corrected two-dimensional cross-sections are combined to form a three-dimensional point cloud model of the tunnel curvature area.