Subway tunnel secondary lining section limit measuring method based on SLAM mobile scanning

Abstract

The invention relates to the field of measurement, in particular to a subway tunnel secondary lining section limit measurement method based on SLAM mobile scanning, which comprises the following steps: step 1, tunnel point cloud acquisition and target clustering point cloud acquisition; step 2, CPIII control point coordinate extraction; step 3, point cloud coordinate conversion; 4, performing point cloud processing, and removing noisy points and non-secondary-lining target point cloud data in the tunnel secondary-lining point cloud; and step 5, performing three-dimensional limit deviation analysis, performing spatial comparative analysis on the point cloud under the construction coordinate system and the subway secondary lining tunnel design three-dimensional model by adopting a voxelization deviation mapping algorithm, extracting a continuous deviation region through voxel clustering, and generating a three-dimensional deviation thermodynamic diagram. According to the method, the two-lining tunnel point cloud obtained by SLAM mobile scanning is accurately converted into a construction coordinate system by combining CPIII high-precision control points in the tunnel, rapid, visual and three-dimensional contrastive analysis with design of a three-dimensional model can be achieved, and visual and reliable data support is provided for line and slope adjustment, track laying and defect improvement.

Description

Technical Field

[0001] This invention relates to the field of cross-sectional clearance measurement of secondary lining sections of subway tunnels, specifically to a method for cross-sectional clearance measurement of secondary lining sections of subway tunnels based on SLAM moving scan. Background Technology

[0002] After the secondary lining of the subway tunnel is completed (either through mining or after the assembly of shield / TBM segments), the secondary lining cross-section needs to be measured to calculate and analyze the deviation between the actual construction profile and the design profile, and to extract the cross-sectional areas with larger deviations. This provides a basis for adjusting the track alignment and slope. If the deviation meets the requirements, the track can be laid according to the original design. If the deviation exceeds the allowable value, the design track alignment needs to be adjusted according to the actual excavation situation to ensure the smoothness of the track laying. Currently, the industry commonly uses two measurement methods. The first is to use a cross-section measuring instrument or total station for single-point measurement. This involves collecting multiple spatial points along the tunnel's direction at 5-meter intervals, connecting the collected points on each cross-section sequentially to generate a measured cross-section, and then comparing it with the design cross-section. This method is inefficient and prone to missing sections exceeding limits. The second method uses a station-mounted scanner, based on a "single-station scanning + multi-station stitching" approach, to acquire the tunnel's point cloud information. Then, it extracts the cross-section point cloud data at certain mileage intervals and compares it with the design cross-section. This method requires time-consuming field deployment of targets and station relocation, and the calculations, analysis, and results are presented in two-dimensional form, lacking intuitiveness. With the compression of subway construction cycles and the increasing demand for 3D visualization, the above methods are no longer sufficient to meet the inspection requirements of "fast, full-section, high-precision, and true 3D."

[0003] In recent years, handheld 3D laser scanners based on the SLAM algorithm have demonstrated advantages in spatial measurement, such as no need for GNSS, no need for leveling, scanning while moving, and continuous data. To address these issues, this invention proposes a handheld mobile scanning method based on SLAM to acquire point clouds of subway tunnel secondary lining. Combined with CPIII control point coordinate information, point cloud data in the construction coordinate system is obtained. Finally, the data is compared and analyzed with the design 3D model to obtain the 3D boundary results of the subway tunnel secondary lining section. Summary of the Invention

[0004] To address the problems of low measurement efficiency, limited data format, weak spatial continuity, and insufficient intuitiveness of results in existing technologies, the present invention aims to provide a method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM mobile scanning. This method utilizes a handheld 3D laser scanner based on SLAM principles for walking scanning, obtains the tunnel secondary lining point cloud in the construction coordinate system based on CPIII control points, and combines noise removal algorithms and spatial deviation analysis algorithms to achieve continuous 3D measurement of the tunnel secondary lining section and extraction of deviation areas, providing intuitive 3D results.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan, comprising the following steps: Step 1: Tunnel point cloud acquisition and target clustering point cloud acquisition. A high-reflectivity circular black and white target is used instead of the prism used for total station measurement. It is inserted into the connecting rod of the CPIII control point and the travel speed is controlled. A 3D laser scanner is used to acquire the 3D point cloud of the inner wall of the secondary lining of the tunnel in real time, and the target forms a high-density, high-reflectivity clustered point cloud in the point cloud. Step 2: Extract CPIII control point coordinates, perform planar fitting on the clustered point cloud, and identify and fit two straight lines on the black and white target that represent the boundary between the black and white regions. The intersection of the two straight lines is used as the three-dimensional coordinate value of the target center. Step 3: Point cloud coordinate transformation. Using the coordinates of the CPIII control points as the control reference, the moving scan point cloud is registered to the absolute coordinate system to obtain the tunnel secondary lining point cloud in the construction coordinate system. Step 4: Point cloud processing, combining automated filtering algorithms with manual intervention, to remove noise and non-secondary lining target point cloud data from the tunnel secondary lining point cloud; Step 5: Three-dimensional boundary deviation analysis. Using the voxelized deviation mapping algorithm, a spatial comparison analysis is performed between the point cloud in the construction coordinate system and the three-dimensional design model of the subway secondary lining tunnel. Continuous deviation regions are extracted through voxel clustering to generate a three-dimensional deviation heat map.

[0006] The above-mentioned method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan, step 1 includes: Step 1-1: Collect CPIII control point data. After the construction of the secondary lining of the subway tunnel, CPIII control points are set up in pairs every 60 m to 120 m along the tunnel direction. The three-dimensional coordinates of the CPIII control points are determined by the free stationing method of total station. Steps 1-2: Field SLAM point cloud scanning. Adjust the target angle and use a handheld SLAM 3D laser scanner to walk at a constant speed along the tunnel. The scanner acquires the 3D point cloud of the inner wall of the tunnel lining in real time. During the walking process, pause every time you pass the CPIII control point to make the target form a high-density, high-reflectivity clustered point cloud in the point cloud.

[0007] The above-mentioned method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM mobile scanning involves dividing the target surface into four right-angled sector regions, with the intervals set as high-reflectivity white and low-reflectivity black. The dividing lines between the black and white regions are two mutually perpendicular straight lines, and their intersection is the target center. The center of the target is in the same spatial position as the center of the total station's standard prism.

[0008] The above-mentioned SLAM-based moving scan method for measuring the clearance of secondary lining sections of subway tunnels, step 2 includes: Step 2-1: Perform intensity filtering on the SLAM point cloud, set the value of the intensity higher than the background environment but lower than the intensity of the white target area as the threshold, retain the point cloud clusters with reflection intensity higher than the set threshold, and use the Euclidean clustering algorithm to cluster the high-intensity point cloud to identify the point cloud cluster corresponding to each target. Step 2-2: Perform plane fitting for each cluster point cloud, and use the RANSAC algorithm to fit the best plane model. Let the fitted plane equation be: ,in, It is a plane normal vector. Let (x, y, z) be a constant term, representing the three-dimensional spatial coordinates of any point on this plane; Steps 2-3: Further project the clustered point cloud onto the fitted plane, transforming the 3D point cloud into a 2D point set. During projection, retain the intensity value of each point. Based on the intensity values, divide each target point cloud into two subsets: a high-intensity point set and a low-intensity point set. and low-intensity point sets It classifies based on a preset intensity threshold and identifies the boundary points between high-intensity and low-intensity regions; Steps 2-4: For the identified boundary points, use a straight line fitting algorithm to fit two straight lines representing the boundary between the black and white regions. and ,straight line The equation is: ,straight line The equation is: ,in, , Indicates the slope. , Represents a constant; Steps 2-4: Coordinates of the target center on the two-dimensional plane That is, the intersection point of the two fitted lines. Solve the system of equations: The coordinates of the center of the circle can be obtained as follows: The calculated two-dimensional center coordinates are back-projected back onto the spatial plane fitted by the formula of the plane equation to obtain the coordinates of the target center in the relative coordinate system in the original SLAM scan point cloud.

[0009] The above-mentioned SLAM-based moving scan method for measuring the clearance of secondary lining sections of subway tunnels, in step 3, assumes that the coordinates of N CPIII control points in the construction coordinate system measured by the total station in step 1 are... The target center coordinates obtained by fitting the SLAM point cloud in step 2 are: By solving a rigid body transformation through corresponding point pairs, the SLAM point cloud in the relative coordinate system is transformed into the construction coordinate system.

[0010] The above-mentioned SLAM-based mobile scanning method for measuring the clearance of secondary lining sections in subway tunnels, step 3 includes: Step 3-1: Calculate the centers of the two point sets: , ,in, This represents the center coordinates of the target point set in the SLAM point cloud. Indicates the center coordinates of the target point set under the construction coordinate system; Step 3-2: Calculate the decentralized point set: , ,in, This represents the new set of target center coordinates formed after decentralization in the SLAM point cloud. This represents the new set of target center coordinate points formed after decentering the construction coordinate system; Step 3-3: Construct the covariance matrix: Perform SVD decomposition on the covariance matrix H: The rotation matrix R is obtained as follows: The translation vector t is: ; Steps 3-4: For each point P in the SLAM point cloud, its coordinates Q in the construction coordinate system are: .

[0011] The above-mentioned SLAM-based mobile scanning method for measuring the clearance of secondary lining sections in subway tunnels, in step 4, employs a statistical outlier removal algorithm to remove discrete noise points in the point cloud and calculates the clearance for each point in the point cloud. Until recently Average distance of neighboring points : If a certain point Greater than the preset threshold If the point is an outlier, it is determined to be an outlier and removed. express The nearest point; For large objects in the point cloud that are clearly not part of the secondary lining structure, they can be manually selected and deleted in the point cloud software, or filtered based on the design cross-section model, and points that deviate from the design model beyond a set threshold can be treated as noise and removed.

[0012] The above-mentioned method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan, step 5 includes: Step 5-1: Convert the point cloud and the design model into voxel meshes of the same resolution. Let the point sets of the measured point cloud and the voxelized mesh of the design model be respectively... and ; Step 5-2: For any voxel grid in the measured point cloud The process iterates through the voxelized mesh of the design model until the voxel mesh of the design model is found. Obtain a voxel mesh and minimum distance between centers This represents the deviation between the measured point cloud at the voxel mesh and the design model. Set a deviation threshold G_max. When the deviation value is greater than G_max, mark the voxel mesh. This is the deviation area; Step 5-3: Extract continuous deviation regions through voxel clustering and generate a three-dimensional deviation heatmap.

[0013] The beneficial effects of this invention, a SLAM-based mobile scanning method for measuring the clearance of secondary lining sections in subway tunnels, are as follows: This invention continuously scans the secondary lining tunnel using a handheld SLAM 3D laser scanner. Based on the CPIII control points within the tunnel, it obtains a point cloud of the secondary lining tunnel in the construction coordinate system. Combined with a 3D design model, it analyzes and calculates the clearance conditions of the secondary lining tunnel, providing data support for subway alignment adjustments, slope adjustments, and track laying. Compared to traditional clearance measurement methods, this invention allows for single-person operation, significantly improving field efficiency. The results overcome the limitations of traditional "two-dimensional cross-sections," providing 3D clearance measurement results, facilitating collaborative decision-making among design, construction, and operation parties. Attached Figure Description

[0014] Figure 1 These are images of the CPIII control point locations during the data acquisition process of this invention. Figure 2 These are images of CPIII points during field SLAM point cloud scanning for this invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be described below in conjunction with specific embodiments and accompanying drawings.

[0016] Example 1 In view of the narrow spatial characteristics of subway tunnels, as well as the complex point cloud denoising and low coordinate transformation accuracy, this embodiment proposes a set of point cloud denoising, high-precision extraction of control points and coordinate registration methods suitable for tunnel environments.

[0017] The method of measuring the cross-sectional clearance of the secondary lining of subway tunnels using SLAM (Simultaneous Localization and Mapping) mobile 3D laser scanning, combined with the CPIII high-precision control points within the tunnel, accurately obtains the point cloud data of the secondary lining tunnel in the construction coordinate system. This enables rapid, intuitive, and 3D visualization comparison and analysis with the design 3D model, providing intuitive and reliable data support for alignment and slope adjustment, track laying, and defect remediation.

[0018] Specifically, a method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM (Simultaneous Localization and Mapping) scanning includes at least the following steps: 1. CPIII Control Point Data Acquisition. After the secondary lining construction of the subway tunnel, CPIII control points are usually set up in pairs every 60 m to 120 m along the tunnel direction, and their three-dimensional coordinates are determined using the free stationing method with a total station.

[0019] 2. Field SLAM Point Cloud Scanning. Before the operation, a high-reflectivity circular black and white target was used instead of the prism for total station surveying. It was inserted into the connecting rod of the CPIII control point. A handheld SLAM 3D laser scanner was used to walk along the tunnel at a constant speed. The scanner acquired the 3D point cloud of the inner wall of the tunnel lining in real time. During the walk, a short pause was made every time the CPIII control point was passed to make the target form a high-density, high-reflectivity cluster in the point cloud.

[0020] 3. CPIII Control Point Coordinate Extraction. Plane fitting is performed on the clustered point cloud, and two straight lines representing the boundary between black and white regions on the black and white target are identified and fitted. The intersection of these two lines is used as the three-dimensional coordinate value of the target center.

[0021] 4. Point cloud coordinate transformation. Using the coordinates of the CPIII control points as the control reference, the moving scan point cloud is registered to the absolute coordinate system to obtain the tunnel secondary lining point cloud in the construction coordinate system.

[0022] 5. SLAM point cloud processing. Combining automated filtering algorithms with manual intervention, noise points and non-secondary lining target point cloud data are removed from the tunnel secondary lining point cloud.

[0023] 6. Three-dimensional boundary deviation analysis. A voxelization deviation mapping algorithm is used to perform spatial comparison analysis between the point cloud in the construction coordinate system and the three-dimensional design model of the subway secondary lining tunnel. The measured point cloud and the design model are voxelized into voxel meshes of the same resolution (e.g., 0.01m × 0.01m × 0.01m). For any voxel mesh in the measured point cloud... The process iterates through the voxelized mesh of the design model until the voxel mesh of the design model is found. to make voxel grid and The center distance is minimized, and this minimum distance is the voxel mesh. The deviation values ​​between the measured point cloud and the design model are calculated, and a deviation threshold G_max is set. When the deviation value is greater than G_max, the voxel is marked as a deviation region. Finally, continuous deviation regions are extracted through voxel clustering, and a three-dimensional deviation heatmap is generated. This algorithm innovatively introduces a voxelized mesh deviation analysis algorithm, realizing continuous three-dimensional deviation analysis of the entire tunnel, avoiding the limitations of traditional discrete cross-sections.

[0024] Example 2 A method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM (Simultaneous Localization and Mapping) mobile scanning includes the following steps.

[0025] Step 1: CPIII Control Point Data Acquisition. After the secondary lining construction of the subway tunnel, CPIII control points are typically set up in pairs every 60 m to 120 m along the tunnel direction, such as... Figure 1 As shown, during the measurement, the CPIII point was inserted by a connecting rod, and the other end of the connecting rod was connected to the standard prism of the total station. The three-dimensional coordinates were determined by the free setting method of the total station, and a total of N CPIII points were measured.

[0026] Step 2: Field SLAM Point Cloud Scanning. Before scanning, insert the end of the CPIII control point connector into a high-reflectivity circular black-and-white target. The target surface is divided into four right-angled sector regions. Two opposing sectors are high-reflectivity white, and the other two are low-reflectivity black. The boundaries between the black and white regions are two perpendicular straight lines, and their intersection is the target center. Figure 2 As shown. The center of the black and white target is in the same spatial position as the center of the total station's standard prism, meaning the 3D coordinates of the prism obtained in step 1 are the 3D coordinates of the center of the circular black and white target. Adjust the target angle to ensure that the handheld scanner can identify targets at an angle while walking. Using a handheld SLAM scanner integrating a 32-line LiDAR, IMU, and panoramic camera, walk slowly and uniformly along the tunnel, and the scanner acquires the 3D point cloud of the tunnel wall in real time. During the walk, pause briefly every time a pair of CPIII control points are passed to allow the target to form a high-density, high-reflectivity cluster in the point cloud.

[0027] Step 3: CPIII Control Point Coordinate Extraction. First, intensity filtering is applied to the SLAM point cloud, retaining point cloud clusters with reflection intensity higher than a set threshold. This threshold should be set to a value higher than the background environment intensity but lower than the intensity of the white target area. Next, Euclidean clustering algorithm is used to cluster the high-intensity point cloud, identifying the point cloud cluster corresponding to each target.

[0028] For each cluster of point clouds, a plane fit is performed, and the RANSAC (Random Sample Consensus) algorithm is used to fit the best plane model. Let the fitted plane equation be: (1), In equation (1), It is a plane normal vector. Let (x, y, z) be a constant term, and let (x, y, z) represent the three-dimensional spatial coordinates of any point on this plane.

[0029] On the fitted plane, the clustered point cloud is further projected, transforming the 3D point cloud into a 2D point set by projecting it onto the fitted plane, while retaining the intensity value of each point during the projection process. On the projected 2D point set, each target point cloud is divided into two subsets based on its intensity value: a high-intensity point set and a low-intensity point set. (Corresponding to the white area) and low-intensity point set (corresponding to the black area), and classify based on a preset intensity threshold to identify the boundary points between high-intensity and low-intensity areas.

[0030] For the identified boundary points, a straight line fitting algorithm is used to fit two straight lines representing the boundary between the black and white regions. and .straight line The equation is: ,straight line The equation is: ,in, , Indicates the slope. , Represents a constant.

[0031] Considering the orthogonal nature of the target design, these two lines should satisfy a perpendicular relationship, that is... This constraint is added during the fitting process to improve the accuracy and robustness of the fitting. , and , The parameters represent the equations of two plane lines. , The geometric meaning of x is the "inclination" of a line relative to the positive x-axis. , The geometric meaning of x and y is the ordinate of the intersection point of the line and the y-axis. x and y represent the variables of the equation. They are universal and both represent the unknowns in the equation. Here, they represent the plane coordinates of any point on the equation of the plane line.

[0032] Coordinates of the target center on a two-dimensional plane That is, the intersection point of the two fitted lines. Solve the system of equations: (2), The coordinates of the center of the circle can be obtained: (3).

[0033] The calculated two-dimensional center coordinates are back-projected back onto the spatial plane fitted by formula (1) to obtain its coordinates in the relative coordinate system in the original SLAM scan point cloud.

[0034] Step 4: Point Cloud Coordinate Transformation. Let the coordinates of the N CPIII control points obtained in Step 1 using a total station in the construction coordinate system be... The target center coordinates obtained by fitting the SLAM point cloud in step 3 are: By using these corresponding point pairs, an optimal rigid body transformation (rotation matrix R and translation vector t) is solved, so that the SLAM point cloud in the relative coordinate system is transformed into the construction coordinate system.

[0035] First, calculate the centers of the two point sets: , (4), of which, This represents the center coordinates of the target point set in the SLAM point cloud. This represents the center coordinates of the target point set in the construction coordinate system.

[0036] Calculate the decentralized point set: , (5), among which, This represents the new set of target center coordinates formed after decentralization in the SLAM point cloud. This represents the new set of target center coordinate points formed after decentering the construction coordinate system, i.e.: Indicates and These represent the new target center coordinate point sets formed by subtracting the point set center coordinates in formula (4) from the target center coordinates in the relative coordinate system of the SLAM point cloud and the construction coordinate system, respectively.

[0037] Construct the covariance matrix: (6).

[0038] Perform SVD decomposition on H: (7).

[0039] Then the rotation matrix R is: (8).

[0040] The translation vector t is: (9), where U, V, and Σ are matrices obtained by performing singular value decomposition (SVD) on the covariance matrix H, where U and V are orthogonal matrices and Σ is a diagonal matrix. and It is the transpose of the corresponding matrix.

[0041] For each point P in the SLAM point cloud, its coordinates Q in the construction coordinate system are: (10).

[0042] Step 5: SLAM point cloud processing. A combination of automated filtering algorithms and manual intervention is used to remove noise points and non-secondary lining target point cloud data from the tunnel secondary lining point cloud.

[0043] A statistical outlier removal algorithm is used to remove discrete noise points from the point cloud. This is applied to each point in the point cloud. Calculate its nearest Average distance of neighboring points : (11), among which, express The nearest point, i.e., the distance point Any one of the k most recent points.

[0044] If a certain point Greater than the preset threshold If the value is not found, the point is considered an outlier and removed. (Threshold value) It is usually set to 1cm.

[0045] For large objects in the point cloud that are clearly not part of the secondary lining structure (such as construction equipment, pipelines, temporary supports, etc.), they can be manually selected and deleted in the point cloud software, or filtered based on the design cross-section model, and points that deviate from the design model by more than a certain threshold can be regarded as noise and removed.

[0046] Step 6: 3D Boundary Deviation Analysis. A voxelization deviation mapping algorithm is used to perform a spatial comparison analysis between the point cloud in the construction coordinate system and the 3D design model of the subway secondary lining tunnel. First, the point cloud and the design model are voxelized into voxel meshes of the same resolution (e.g., 0.01m × 0.01m × 0.01m). Let the measured point cloud and the voxelized mesh point sets of the design model be respectively... and For any voxel grid in the measured point cloud The process iterates through the voxelized mesh of the design model until the voxel mesh of the design model is found. to make voxel grid and The center distance is the smallest, and this minimum distance That is, voxel grid The deviation between the measured point cloud and the design model at the location. (12).

[0047] Set a deviation threshold G_max, which is typically set to 5-8cm. In this embodiment, it is set to 5cm. When the deviation value... When the value is greater than G_max, the voxel is marked as a deviation region; finally, continuous deviation regions are extracted by voxel clustering and a three-dimensional deviation heatmap is generated.

[0048] Cluster analysis was performed on the voxel mesh marked as deviation regions. A connectivity-based clustering algorithm was used to aggregate voxels that are spatially adjacent and whose deviation values ​​all exceed the threshold G_max into a continuous deviation region. Each cluster region represents a tunnel section where the actual construction profile deviates significantly from the design model.

[0049] Each voxel grid is assigned a corresponding color value based on its deviation value; the larger the deviation, the darker the color (e.g., dark red), and the smaller the deviation, the lighter the color (e.g., light blue), forming a 3D heatmap with a color gradient. This heatmap can be overlaid on the original point cloud or design model for visualization, intuitively reflecting the spatial distribution of clearance deviations throughout the tunnel.

[0050] In the 3D deviation heatmap, the distribution of deviations throughout the tunnel space is intuitively visualized through color gradients (e.g., from blue representing small deviations to red representing large deviations). This helps engineers quickly identify the location, extent, and severity of sections exceeding limits, thus providing a reliable basis for decision-making regarding subway tunnel alignment and gradient adjustments, track laying, and defect remediation. It avoids the limitations of traditional two-dimensional analysis and achieves more efficient 3D spatial assessment. This algorithm innovatively introduces a voxelized mesh deviation analysis algorithm, realizing continuous three-dimensional deviation analysis of the entire tunnel, thus avoiding the limitations of traditional discrete cross-sections.

[0051] The field efficiency of this invention is significantly better than that of the existing method for measuring the clearance of the secondary lining section of subway tunnels. It can be completed by a single person with a single instrument. By using a three-dimensional spatial overall comparison algorithm, the measured point cloud and the design model are analyzed for spatial differences, and the three-dimensional results of the over-limit section are identified and output, replacing the traditional "two-dimensional section + mileage list" mode.

[0052] The above embodiments are merely illustrative of the structural concept and features of the present invention, intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly, and should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan, characterized in that, Includes the following steps: Step 1: Tunnel point cloud acquisition and target clustering point cloud acquisition. A high-reflectivity circular black and white target is used instead of the prism used for total station measurement. It is inserted into the connecting rod of the CPIII control point and the travel speed is controlled. A 3D laser scanner is used to acquire the 3D point cloud of the inner wall of the secondary lining of the tunnel in real time, and the target forms a high-density, high-reflectivity clustered point cloud in the point cloud. Step 2: Extract CPIII control point coordinates, perform planar fitting on the clustered point cloud, and identify and fit two straight lines on the black and white target that represent the boundary between the black and white regions. The intersection of the two straight lines is used as the three-dimensional coordinate value of the target center. Step 3: Point cloud coordinate transformation. Using the coordinates of the CPIII control points as the control reference, the moving scan point cloud is registered to the absolute coordinate system to obtain the tunnel secondary lining point cloud in the construction coordinate system. Step 4: Point cloud processing, combining automated filtering algorithms with manual intervention, to remove noise and non-secondary lining target point cloud data from the tunnel secondary lining point cloud; Step 5: Three-dimensional boundary deviation analysis. Using the voxelized deviation mapping algorithm, a spatial comparison analysis is performed between the point cloud in the construction coordinate system and the three-dimensional design model of the subway secondary lining tunnel. Continuous deviation regions are extracted through voxel clustering to generate a three-dimensional deviation heat map.

2. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 1, characterized in that, Step 1 includes: Step 1-1: Collect CPIII control point data. After the construction of the secondary lining of the subway tunnel, CPIII control points are set up in pairs every 60 m to 120 m along the tunnel direction. The three-dimensional coordinates of the CPIII control points are determined by the free stationing method of total station. Steps 1-2: Field SLAM point cloud scanning. Adjust the target angle and use a handheld SLAM 3D laser scanner to walk at a constant speed along the tunnel. The scanner acquires the 3D point cloud of the inner wall of the tunnel lining in real time. During the walking process, pause every time you pass the CPIII control point to make the target form a high-density, high-reflectivity clustered point cloud in the point cloud.

3. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 2, characterized in that, The target surface is divided into four right-angled sector regions, with the intervals set as high-reflectivity white and low-reflectivity black. The dividing lines between the black and white regions are two mutually perpendicular straight lines, and their intersection is the target center. The center of the target is in the same spatial position as the center of the total station's standard prism.

4. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 1, characterized in that, Step 2 includes: Step 2-1: Perform intensity filtering on the SLAM point cloud, set the value of the intensity higher than the background environment but lower than the intensity of the white target area as the threshold, retain the point cloud clusters with reflection intensity higher than the set threshold, and use the Euclidean clustering algorithm to cluster the high-intensity point cloud to identify the point cloud cluster corresponding to each target. Step 2-2: Perform plane fitting for each cluster point cloud, and use the RANSAC algorithm to fit the best plane model. Let the fitted plane equation be: ,in, It is a plane normal vector. Let (x, y, z) be a constant term, representing the three-dimensional spatial coordinates of any point on this plane; Steps 2-3: Further project the clustered point cloud onto the fitted plane, transforming the 3D point cloud into a 2D point set. During projection, retain the intensity value of each point. Based on the intensity values, divide each target point cloud into two subsets: a high-intensity point set and a low-intensity point set. and low-intensity point sets It classifies based on a preset intensity threshold and identifies the boundary points between high-intensity and low-intensity regions; Steps 2-4: For the identified boundary points, use a straight line fitting algorithm to fit two straight lines representing the boundary between the black and white regions. and ,straight line The equation is: ,straight line The equation is: ,in, , Indicates the slope. , Represents a constant; Steps 2-4: Coordinates of the target center on the two-dimensional plane That is, the intersection point of the two fitted lines. Solve the system of equations: The coordinates of the center of the circle can be obtained as follows: The calculated two-dimensional center coordinates are back-projected back onto the spatial plane fitted by the formula of the plane equation to obtain the coordinates of the target center in the relative coordinate system in the original SLAM scan point cloud.

5. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 1, characterized in that, In step 3, let the coordinates of the N CPIII control points in the construction coordinate system measured by the total station in step 1 be... The target center coordinates obtained by fitting the SLAM point cloud in step 2 are: By solving a rigid body transformation through corresponding point pairs, the SLAM point cloud in the relative coordinate system is transformed into the construction coordinate system.

6. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 4, characterized in that, Step 3 includes: Step 3-1: Calculate the centers of the two point sets: , ,in, This represents the center coordinates of the target point set in the SLAM point cloud. Indicates the center coordinates of the target point set under the construction coordinate system; Step 3-2: Calculate the decentralized point set: , ,in, This represents the new set of target center coordinates formed after decentralization in the SLAM point cloud. This represents the new set of target center coordinate points formed after decentering the construction coordinate system; Step 3-3: Construct the covariance matrix: Perform SVD decomposition on the covariance matrix H: The rotation matrix R is obtained as follows: The translation vector t is: ; Steps 3-4: For each point P in the SLAM point cloud, its coordinates Q in the construction coordinate system are: .

7. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 1, characterized in that, In step 4, a statistical outlier removal algorithm is used to remove discrete noise points from the point cloud, and the algorithm is used to calculate the number of outliers for each point in the point cloud. Until recently Average distance of neighboring points : If a certain point Greater than the preset threshold If the point is an outlier, it is determined to be an outlier and removed. express The nearest point; For large objects in the point cloud that are clearly not part of the secondary lining structure, they can be manually selected and deleted in the point cloud software, or filtered based on the design cross-section model, and points that deviate from the design model beyond a set threshold can be treated as noise and removed.

8. The method for measuring the clearance of secondary lining sections of subway tunnels based on SLAM moving scan according to claim 1, characterized in that, Step 5 includes: Step 5-1: Convert the point cloud and the design model into voxel meshes of the same resolution. Let the point sets of the measured point cloud and the voxelized mesh of the design model be respectively... and ; Step 5-2: For any voxel grid in the measured point cloud The process iterates through the voxelized mesh of the design model until the voxel mesh of the design model is found. Obtain a voxel mesh and minimum distance between centers This represents the deviation between the measured point cloud at the voxel mesh and the design model. Set a deviation threshold G_max. When the deviation value is greater than G_max, mark the voxel mesh. This is the deviation area; Step 5-3: Extract continuous deviation regions through voxel clustering and generate a three-dimensional deviation heatmap.

Images