Tunnel profile change detection method, device and storage medium based on mobile laser scanning

Abstract

The present invention relates to a method, device, and storage medium for detecting tunnel contour changes based on mobile laser scanning. The method comprises: obtaining laser point clouds of each section within the tunnel using mobile laser scanning; obtaining multiple feature points annotated based on the section point clouds at the starting point; segmenting the tunnel based on neighborhood density, dividing the tunnel into multiple tunnel segments based on actual demarcation points, obtaining the length and contour of each tunnel segment, and combining the lengths and contours of all tunnel segments to obtain the tunnel contour. Compared with existing technologies, the present invention has the advantages of improving the accuracy and speed of contour detection for complex tunnels.

Description

Technical Field

[0001] The present invention relates to the field of tunnel profile detection, and in particular to a tunnel profile change detection method, device and storage medium based on mobile laser scanning. Background Art

[0002] For tunnels with complex geometries, a single section may encompass multiple cross-section types. In the absence of design data or difficulty obtaining it, there is an urgent need to accurately determine the demarcation points where tunnel contours change and further extract the contours of each segment. This process is of great practical significance for the production of point cloud orthophotos and the reconstruction of existing tunnel models.

[0003] In the existing technology, depending on the available hardware equipment and data processing methods, the commonly used technologies for identifying and extracting tunnel profile changes include measurement robotics, photogrammetry, and 3D laser scanning, as follows:

[0004] (1) The measurement robot technology first determines the location of each segment through manual inspection, and then uses the measurement robot to measure the tunnel section to obtain the tunnel profile. This technology has high accuracy, but low efficiency and high labor costs. In addition, the data obtained only represents the current local area, and the measurement results are limited.

[0005] (2) Photogrammetry technology takes photos of the interior of the tunnel and uses image processing methods to determine changes in the tunnel contour. However, this technology has high requirements for lighting, and the image data lacks depth information, making it difficult to obtain spatial geometric information, which is not conducive to extracting the tunnel contour.

[0006] (3) Three-dimensional laser scanning technology has been used in tunnel contour extraction because it can obtain high-precision, high-density three-dimensional point cloud data inside the tunnel. However, the existing methods are highly dependent on the central axis of the tunnel and are easily interfered by factors such as occlusion of facilities and equipment.

[0007] In summary, the existing related technologies can improve the efficiency of tunnel contour extraction to a certain extent, but the efficiency of tunnel contour change recognition is generally not high. Tunnel contour change recognition and extraction technology still faces some challenges in practical application. Summary of the Invention

[0008] The purpose of the present invention is to provide a tunnel profile change detection method, device and storage medium based on mobile laser scanning.

[0009] The purpose of the present invention can be achieved by the following technical solutions:

[0010] A tunnel profile change detection method based on mobile laser scanning, comprising:

[0011] Step S1: Acquire laser point clouds of each section in the tunnel based on mobile laser scanning;

[0012] Step S2: filtering the acquired laser point cloud;

[0013] Step S3: obtaining multiple feature points based on the cross-section point cloud annotation at the starting point;

[0014] Step S4: Determine whether there is a next section. If yes, calculate the neighborhood density of each feature point in the laser point cloud of the next section. Otherwise, execute step S9.

[0015] Step S5: Determine whether there is at least one feature point whose neighborhood density is less than a pre-configured threshold. If yes, execute step S6; otherwise, return to step S4.

[0016] Step S6: taking the current section as a potential dividing point;

[0017] Step S7: Calculate the neighborhood density of each feature point in the laser point cloud of the next section, and determine whether there are still feature points whose neighborhood density is less than the pre-configured threshold. If so, execute step S8; otherwise, return to step S4.

[0018] Step S8: Determine whether the length from the current section to the potential dividing point exceeds a pre-configured threshold length. If so, use the potential dividing point as the actual dividing point, return to the actual dividing point, and obtain multiple feature points based on the actual dividing point to replace the original feature points, and return to step S4;

[0019] Step S9: Divide the tunnel into multiple tunnel segments based on the actual dividing points, obtain the length and profile of each tunnel segment, and combine the lengths and profiles of all tunnel segments to obtain the profile of the tunnel.

[0020] The filtering method in step S2 adopts straight-through filtering and spatial statistical filtering.

[0021] The neighborhood density of the feature point is specifically the number of laser points in the cross section whose distance from the feature point is less than a first set threshold distance.

[0022] The first set threshold distance is 20 centimeters.

[0023] The preconfigured threshold length is the maximum length of the equipment installed on the inner wall of the tunnel along the axial direction of the tunnel.

[0024] In the process of obtaining the length and profile of each segmented tunnel, the length of the segmented tunnel is calibrated based on the milepost in the tunnel.

[0025] In the process of obtaining the length and outline of each segmented tunnel, the process of obtaining the outline of the segmented tunnel specifically includes:

[0026] Step S9-1: Selecting point clouds of multiple sections within the segmented tunnel section;

[0027] Step S9-2: Projecting the selected point cloud along a direction parallel to the axis of the segmented tunnel onto a plane perpendicular to the axis of the segmented tunnel to obtain a cross-sectional projection diagram;

[0028] Step S9-3: Select any two projection points in the cross-sectional projection diagram;

[0029] Step S9-4: Draw two circles with a pre-configured radius through the two selected projection points. If no other projection points exist within any of the circles, the two selected projection points are used as boundary points.

[0030] Step S9-5: Determine whether all projection point combinations have been traversed. If so, execute step S9-6. Otherwise, select two other projection point combinations that have not been traversed and return to step S9-4.

[0031] Step S9-6: Connect the boundary points in sequence in a clockwise or counterclockwise direction to obtain the outline of the segmented tunnel.

[0032] In step S9-1, the selected multiple sections are distributed at equal distances.

[0033] A tunnel profile change detection device based on mobile laser scanning includes a memory, a processor, and a program stored in the memory, characterized in that the processor implements the above-mentioned method when executing the program.

[0034] A storage medium stores a program thereon, wherein the program implements the above method when executed.

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

[0036] 1. For complex-shaped tunnels, contour changes are judged based on the neighborhood features of cross-section feature points, which can efficiently segment the tunnel according to contour shape. In addition, the neighborhood radius and judgment threshold are set specifically to account for interference such as obstruction by facilities and equipment in the tunnel, effectively avoiding misjudgment of tunnel contour changes caused by related interference factors.

[0037] 2. Combined with the boundary extraction algorithm, the tunnel contour can be extracted without the tunnel centerline. The point cloud projection of multiple sections within the tunnel segment is used as the basic data for contour extraction, avoiding errors that may be caused by equipment occlusion in a single section. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 Schematic diagram of the main steps of the method of the present invention;

[0039] Figure 2 This is an example diagram for setting the first cross-section feature point;

[0040] Figure 3 This is a schematic diagram of the tunnel cross-section point cloud;

[0041] Figure 4 for Figure 3 The contour extraction results of the cross-section point cloud are shown;

[0042] Figure 5 Extract the rendering of a tunnel outline;

[0043] Figure 6 This is a local model diagram of a tunnel where the contour changes. DETAILED DESCRIPTION

[0044] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0045] A tunnel profile change detection method based on mobile laser scanning, such as Figure 1 Shown, including:

[0046] Step S1: Acquire laser point clouds of each section in the tunnel based on mobile laser scanning;

[0047] Step S2: filtering the acquired laser point cloud;

[0048] During the scanning process, due to the influence of observation conditions, the instrument itself and external environmental conditions, the collected raw point cloud data often contains noise points, so filtering is required. The filtering methods used are straight-through filtering and spatial statistical filtering. Among them, straight-through filtering is used to remove noise points outside the tunnel. Its threshold parameters can be determined after a preliminary estimate of the geometric dimensions of the tunnel; spatial statistical filtering is mainly used to remove discrete noise points. Its basic principle is to perform K-neighborhood statistical analysis on each point, calculate the average distance from the point to its K neighboring points, and assume that the result obeys a Gaussian distribution. The shape of the Gaussian distribution depends on the mean and standard deviation. Points with an average distance outside the given threshold range are removed. This processing method can effectively remove noise points from the tunnel section point cloud. The specific process is as follows:

[0049] (1) Input the cross-section point cloud data to be denoised;

[0050] (2) According to the two coordinate axis directions of the cross-section coordinate system where the point cloud is located, the tunnel contour threshold is determined based on prior knowledge, and the point cloud outside the threshold is eliminated;

[0051] (3) Calculate the average distance between neighboring points in the K neighborhood of each point in the remaining point cloud;

[0052] (4) Calculate the mean and standard deviation of the average distance corresponding to each point;

[0053] (5) Taking 1 times the mean square error as the threshold, remove the points whose average distance to the adjacent points is greater than the threshold to obtain the denoised point cloud. If the denoising effect is poor, the corresponding K value and threshold can be adjusted.

[0054] Step S3: obtaining multiple feature points based on the cross-section point cloud annotation at the starting point;

[0055] In this embodiment, the feature points are obtained by manual setting. Of course, in other embodiments, other calculation or data analysis methods may be used for determination.

[0056] Step S4: Determine whether there is a next section. If yes, calculate the neighborhood density of each feature point in the laser point cloud of the next section. Otherwise, execute step S9.

[0057] In this embodiment, the neighborhood density of the feature point is specifically the number of laser points in the cross section whose distance from the feature point is less than a first set threshold distance.

[0058] Step S5: Determine whether there is at least one feature point whose neighborhood density is less than a pre-configured threshold. If yes, execute step S6; otherwise, return to step S4.

[0059] In this embodiment, the first set threshold distance is 20 centimeters. In other embodiments, it can also be set according to actual conditions.

[0060] Step S6: taking the current section as a potential dividing point;

[0061] Step S7: Calculate the neighborhood density of each feature point in the laser point cloud of the next section, and determine whether there are still feature points whose neighborhood density is less than the pre-configured threshold. If so, execute step S8; otherwise, return to step S4.

[0062] The preconfigured threshold value N0 is determined by the density of the point cloud collected by the laser scanner, the angular resolution of the scanner, and the distance from the scanner to the tunnel wall. The specific setting method is as follows:

[0063]

[0064] In the above formula, N0 is the preconfigured threshold position; R is the first set threshold distance, which is set to 20 cm in the present invention; D is the maximum spacing between two adjacent scanning points on the inner wall of the tunnel; L is the maximum distance from the scanner laser emission point to the inner wall of the tunnel; α is the angular resolution of the scanner used.

[0065] Step S8: Determine whether the length from the current section to the potential dividing point exceeds a pre-configured threshold length. If so, use the potential dividing point as the actual dividing point, return to the actual dividing point, and obtain multiple feature points based on the actual dividing point to replace the original feature points, and return to step S4;

[0066] Generally, the preconfigured threshold length is the maximum length of the equipment installed on the inner wall of the tunnel along the tunnel axis, to avoid mis-segmentation caused by some equipment.

[0067] Step S9: Divide the tunnel into multiple tunnel segments based on the actual dividing points, obtain the length and profile of each tunnel segment, and combine the lengths and profiles of all tunnel segments to obtain the profile of the tunnel.

[0068] In the process of obtaining the length and outline of each segmented tunnel, the length of the segmented tunnel is calibrated based on the mileposts in the tunnel. The specific process is as follows:

[0069] In actual operation, an odometer is usually used to record the mileage of the cross-section obtained by mobile laser scanning, and then the mileage markers in the tunnel are used for correction. The mileage correction principle is as follows:

[0070] In a shorter distance, it can be assumed that the mileage measurement error is proportional to the travel distance, that is:

[0071]

[0072] Where: S is the measurement distance, is the actual distance, k is the proportional error coefficient. During calibration, a known reference distance value is used. The error coefficient can be obtained by combining the measured value S with the above formula:

[0073]

[0074] Then the value after any distance correction is:

[0075]

[0076] For shield tunnels, we can also take advantage of the high precision and fixed width of shield segment splicing. By identifying the annular joints, we can determine the number of segments, calculate the actual distance between different annular joints based on the segment width, and then combine the mileage reference distance between the first and last annular joints to achieve mileage correction.

[0077] In addition, in the process of obtaining the length and outline of each segmented tunnel, the process of obtaining the outline of the segmented tunnel specifically includes:

[0078] Step S9-1: Selecting point clouds of multiple sections within the segmented tunnel section;

[0079] Step S9-2: Projecting the selected point cloud along a direction parallel to the axis of the segmented tunnel onto a plane perpendicular to the axis of the segmented tunnel to obtain a cross-sectional projection diagram;

[0080] Step S9-3: Select any two projection points in the cross-sectional projection diagram;

[0081] Step S9-4: Draw two circles with a pre-configured radius through the two selected projection points. If no other projection points exist within any of the circles, the two selected projection points are used as boundary points.

[0082] Step S9-5: Determine whether all projection point combinations have been traversed. If so, execute step S9-6. Otherwise, select two other projection point combinations that have not been traversed and return to step S9-4.

[0083] Step S9-6: Connect the boundary points in sequence in a clockwise or counterclockwise direction to obtain the outline of the segmented tunnel.

[0084] In step S9-1, the selected multiple sections are distributed at equal distances.

[0085] In some embodiments, the boundary points of the tunnel contour are extracted using Matlab's built-in Boundary function with the Alpha Shapes algorithm as the core. Finally, the tunnel contour is obtained by sequentially connecting the boundary points. If the extraction effect is not good, the shrinkage factor of the Boundary function can be appropriately adjusted, such as Figure 3 and Figure 4 Shown are the contour extraction results.

[0086] The method was applied to a tunnel in a certain subway line section. The tunnel in this section is a mining tunnel and contains tunnel profiles of various shapes. Due to the confidentiality agreement, this document sets the start and end mileage range of the tunnel as K0+0.00-K2+331.60 based on the actual distance of the tunnel. The method of this application was used to identify and extract the changes in the tunnel profile, and 16 profiles were obtained in sequence. The profiles extracted from each segmented tunnel are as follows: Figure 5 The starting and ending mileages of each segmented tunnel are shown in Table 1.

[0087] Table 1

[0088]

[0089]

[0090] The tunnel contour change identification and extraction results of this application can be applied to reconstruct tunnel models, such as Figure 6 As shown, this is a local model of a certain contour change in the tunnel.

[0091] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

Claims

1. A tunnel profile change detection method based on mobile laser scanning, characterized in that: include: Step S1: Acquire laser point clouds of each section in the tunnel based on mobile laser scanning; Step S2: filtering the acquired laser point cloud; Step S3: obtaining multiple feature points based on the cross-section point cloud annotation at the starting point; Step S4: Determine whether there is a next section. If yes, calculate the neighborhood density of each feature point in the laser point cloud of the next section. Otherwise, execute step S9. Step S5: Determine whether there is at least one feature point whose neighborhood density is less than a pre-configured threshold. If yes, execute step S6; otherwise, return to step S4. Step S6: taking the current section as a potential dividing point; Step S7: Calculate the neighborhood density of each feature point in the laser point cloud of the next section, and determine whether there are still feature points whose neighborhood density is less than the pre-configured threshold. If so, execute step S8; otherwise, return to step S4. Step S8: Determine whether the length from the current section to the potential dividing point exceeds a pre-configured threshold length. If so, use the potential dividing point as the actual dividing point, return to the actual dividing point, and obtain multiple feature points based on the actual dividing point to replace the original feature points, and return to step S4; Step S9: Divide the tunnel into multiple tunnel segments based on the actual dividing points, obtain the length and profile of each tunnel segment, and combine the lengths and profiles of all tunnel segments to obtain the profile of the tunnel.

2. The tunnel profile change detection method based on mobile laser scanning according to claim 1 is characterized in that: The filtering method in step S2 adopts straight-through filtering and spatial statistical filtering.

3. The tunnel profile change detection method based on mobile laser scanning according to claim 1 is characterized in that: The neighborhood density of the feature point is specifically the number of laser points in the cross section whose distance from the feature point is less than a first set threshold distance.

4. The tunnel profile change detection method based on mobile laser scanning according to claim 3 is characterized in that: The first set threshold distance is 20 centimeters.

5. The tunnel profile change detection method based on mobile laser scanning according to claim 1 is characterized in that: The preconfigured threshold length is the maximum length of the equipment installed on the inner wall of the tunnel along the axial direction of the tunnel.

6. The tunnel profile change detection method based on mobile laser scanning according to claim 1 is characterized in that: In the process of obtaining the length and profile of each segmented tunnel, the length of the segmented tunnel is calibrated based on the milepost in the tunnel.

7. The method for detecting tunnel contour changes based on mobile laser scanning according to claim 1, characterized in that: In the process of obtaining the length and outline of each segmented tunnel, the process of obtaining the outline of the segmented tunnel specifically includes: Step S9-1: Selecting point clouds of multiple sections within the segmented tunnel section; Step S9-2: Projecting the selected point cloud along a direction parallel to the axis of the segmented tunnel onto a plane perpendicular to the axis of the segmented tunnel to obtain a cross-sectional projection diagram; Step S9-3: Select any two projection points in the cross-sectional projection diagram; Step S9-4: Draw two circles with a pre-configured radius through the two selected projection points. If no other projection points exist within any of the circles, the two selected projection points are used as boundary points. Step S9-5: Determine whether all projection point combinations have been traversed. If so, execute step S9-6. Otherwise, select two other projection point combinations that have not been traversed and return to step S9-4. Step S9-6: Connect the boundary points in sequence in a clockwise or counterclockwise direction to obtain the outline of the segmented tunnel.

8. The method for detecting tunnel contour changes based on mobile laser scanning according to claim 7, characterized in that: In step S9-1, the selected multiple sections are distributed at equal distances.

9. A tunnel profile change detection device based on mobile laser scanning, comprising a memory, a processor, and a program stored in the memory, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

10. A storage medium having a program stored thereon, characterized in that: When the program is executed, the method according to any one of claims 1 to 8 is implemented.

Images