Method and System for Point Cloud Segmentation and Construction Progress Information Extraction in Tunnel Secondary Lining

Abstract

This invention proposes a method and system for point cloud segmentation and construction progress information extraction in tunnel secondary lining, relating to the fields of underground engineering and construction quality monitoring. The method includes: calculating the roughness of the tunnel point cloud; assigning roughness values ​​to a voxelized point cloud, extracting points with roughness values ​​meeting set conditions as seed points to form a seed point cloud; traversing the seed point cloud, performing a radius search, and saving the searched points as a new point cloud; deleting duplicate points in the new point cloud; fitting the plane A containing the bottom point cloud to the secondary lining point cloud, and projecting the secondary lining point cloud onto plane A to obtain a planar point cloud; calculating the covariance matrix and linearity values ​​of the planar point cloud, extracting point clouds that meet a linear threshold as linear region point clouds, and performing line extraction; calculating the maximum length of each extracted straight line point cloud, and taking the average of the maximum lengths as the distance for the secondary lining advance. This invention improves the overall construction safety and quality of highway tunnels.

Description

Technical Field

[0001] This invention belongs to the field of underground engineering and construction quality monitoring technology, and particularly relates to a method and system for segmenting point clouds and extracting construction progress information for secondary tunnel lining. Background Technology

[0002] The quality of secondary lining construction is closely related to the overall quality of tunnel construction and is an important link in ensuring the overall quality of tunnel construction. The construction progress of secondary lining is also crucial. Accurate and timely feedback of progress information can help to rationally arrange subsequent procedures and ensure the efficiency of each construction step.

[0003] Secondary lining construction may encounter problems such as misalignment, exposed reinforcement, insufficient lining thickness, and unclear progress information during construction management. The following issues also exist in progress calculation:

[0004] (1) How to obtain a high-completeness lining point cloud;

[0005] (2) Insufficient calculation accuracy.

[0006] These problems can directly or indirectly affect the progress and quality of the project construction, and may even lead to safety accidents. Summary of the Invention

[0007] To overcome the shortcomings of the prior art, this invention provides a method and system for segmenting point clouds and extracting construction progress information for secondary tunnel lining. This method obtains a highly complete lining point cloud, improves calculation accuracy, ensures the construction quality and progress of secondary lining, and enhances the overall construction safety and quality of highway tunnels. It is of great significance for monitoring the secondary lining of tunnels during the construction period and managing the construction progress.

[0008] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions:

[0009] The first aspect of this invention provides a method for segmenting point clouds and extracting construction progress information for secondary tunnel lining.

[0010] The method for segmenting point clouds and extracting construction progress information for secondary tunnel lining includes the following steps:

[0011] The tunnel point cloud is denoised and voxelized. For each point in the voxelized point cloud, a neighborhood plane is fitted and the distance from each point to its neighborhood fitted plane is calculated. The roughness is then calculated.

[0012] The calculated roughness values ​​are assigned to the voxelized point cloud to obtain a point cloud with roughness features. Points with roughness values ​​that meet the set conditions are extracted as seed points to form a seed point cloud.

[0013] Traverse every point in the seed point cloud, perform a radius search on the denoised point cloud, and save the searched points as a new point cloud;

[0014] Delete duplicate points in the new point cloud to obtain a secondary point cloud;

[0015] Fit the point cloud of the secondary lining to obtain the plane A containing the bottom point cloud, and project the secondary lining point cloud onto plane A to obtain the planar point cloud;

[0016] For each point in the planar point cloud, calculate the covariance matrix and the linear value. Extract the point cloud that satisfies the linear threshold as the linear region point cloud, and extract the straight line from the linear region point cloud.

[0017] For the extracted point clouds of multiple straight lines, calculate the maximum length of each point cloud, and take the average of the maximum lengths as the distance for the secondary lining advance.

[0018] The second aspect of this invention provides a system for segmenting point clouds and extracting construction progress information for secondary tunnel lining.

[0019] A point cloud segmentation and construction progress information extraction system for tunnel secondary lining includes:

[0020] The roughness calculation module is configured to: denoise and voxelize the tunnel point cloud, fit a neighborhood plane to each point in the voxelized point cloud and calculate the distance from each point to its neighborhood fitted plane, and calculate the roughness.

[0021] The roughness assignment and filtering module is configured to: assign the calculated roughness value to the voxelized point cloud to obtain a point cloud with roughness features, and extract points whose roughness values ​​meet the set conditions as seed points to form a seed point cloud.

[0022] The radius search module is configured to: traverse every point in the seed point cloud, perform a radius search on the denoised point cloud, and save the searched points as a new point cloud;

[0023] The duplicate point deletion module is configured to delete duplicate points in the new point cloud to obtain a secondary lining point cloud.

[0024] The projection module is configured to: fit the plane A containing the bottom point cloud to the secondary liner point cloud, and project the secondary liner point cloud onto plane A to obtain the planar point cloud;

[0025] The linear region point cloud acquisition module is configured to: calculate the covariance matrix for each point in the planar point cloud and calculate the linear value; extract the point cloud that meets the linear threshold as the linear region point cloud; and extract straight lines from the linear region point cloud.

[0026] The secondary lining advance distance calculation module is configured to: calculate the maximum length of each extracted straight point cloud, and take the average of the maximum lengths as the secondary lining advance distance.

[0027] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the steps in the method for segmenting point clouds and extracting construction progress information for secondary tunnel lining as described in the first aspect of the present invention.

[0028] The fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the method for segmenting point clouds and extracting construction progress information for secondary tunnel lining as described in the first aspect of the present invention.

[0029] The above one or more technical solutions have the following beneficial effects:

[0030] This invention provides a method and system for point cloud segmentation and construction progress information extraction in tunnel secondary lining. The method first performs combined denoising processing on the original tunnel point cloud data to reduce noise within the tunnel and minimize its impact on subsequent processing. Then, the processed point cloud is divided into voxel grids, where the size of the voxel grid is determined by the point cloud density. Roughness is calculated on the voxelized point cloud, where roughness is the distance from a point to a neighboring fitted plane, and an appropriate computational radius is set to calculate the neighborhood roughness. After obtaining the point cloud roughness, points within a specified threshold are extracted. These roughness-calculated points are then used as seed points for secondary extraction in the original point cloud using a kd-tree-based method. The extracted secondary lining point cloud is then projected onto the xoy plane, and linear regions are extracted from the projected point cloud using a PCA-based method. The extracted linear regions are then fitted with multiple straight lines using a least squares-based method. Finally, the lengths of the obtained multiple branches are calculated, and the average value is used as the construction progress distance. This method obtains a highly complete lining point cloud, improves calculation accuracy, reduces computational load, and has good applicability and segmentation integrity for complex construction tunnel scenarios. The error after extracting the advance information is small, and it has good robustness, making it suitable for extracting advance information of tunnel secondary lining.

[0031] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0032] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0033] Figure 1 This is an overall flowchart of Embodiment 1 of the present invention.

[0034] Figure 2 The original point cloud provided in Embodiment 1 of the present invention.

[0035] Figure 3 This is a voxelized point cloud provided in Embodiment 1 of the present invention.

[0036] Figure 4 This is a seed point cloud that meets the roughness requirements provided in Embodiment 1 of the present invention.

[0037] Figure 5 This is a secondary point cloud obtained by secondary segmentation of the seed point cloud, as provided in Embodiment 1 of the present invention.

[0038] Figure 6 This is a diagram illustrating the process of obtaining the advance distance from the secondary lining point cloud, as provided in Embodiment 1 of the present invention. Detailed Implementation

[0039] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0040] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0041] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0042] The overall concept proposed in this invention is as follows:

[0043] To address the current lack of applications for segmentation, progress information extraction, and construction quality inspection of 3D laser point clouds in tunnel secondary lining, this invention provides an algorithm for point cloud segmentation and progress information extraction in tunnel secondary lining. The method is as follows:

[0044] First, the original tunnel point cloud data is combined and denoised to reduce noise inside the tunnel and minimize its impact on subsequent processing.

[0045] Then, the processed point cloud is divided into voxel grids, where the size of the voxel grid is determined by the point cloud density.

[0046] Roughness is calculated on the point cloud after voxelization, where roughness is the distance from a point to the neighborhood fitting plane, and an appropriate calculation radius is set to calculate the neighborhood roughness.

[0047] After obtaining the point cloud roughness, points within a specified threshold are extracted. Then, the points extracted after roughness calculation are used as seed points, and secondary extraction is performed on the original point cloud using a kd-tree-based method. After that, the extracted secondary lining point cloud is projected onto the xoy plane, and linear region extraction is performed on the projected point cloud using a PCA-based method.

[0048] The extracted linear region is fitted with multiple straight lines using a least squares method. Finally, the lengths of the multiple branches are calculated and the average value is used as the advance distance.

[0049] This method reduces computational load and has good applicability and segmentation integrity for complex construction tunnel scenarios. The error after extracting the advance information is small, and it has good robustness. It is suitable for extracting advance information for secondary tunnel lining.

[0050] Example 1

[0051] This embodiment discloses a method for segmenting point clouds and extracting construction progress information for secondary tunnel lining.

[0052] like Figure 1 As shown, the method for segmenting point clouds and extracting construction progress information for secondary tunnel lining includes the following steps:

[0053] Step 1: Use a combination of statistical filtering and radius filtering to denoise the tunnel point cloud P, and obtain the tunnel point cloud P2 with noise and outliers removed;

[0054] Specifically, the denoising steps of radius statistical combined filtering are as follows:

[0055] Traverse the point cloud and remove points whose number of points within the radius does not meet a certain threshold 'a'.

[0056] For a point that satisfies the number of points within its radius, calculate the average distance d1 of its neighboring points within the radius, then calculate the global standard deviation b and the global point cloud average distance d2. If d1 > (b + d2), then remove the point.

[0057] Step 2: Voxelize the denoised point cloud P2 to obtain the voxelized point cloud P3. The size of the voxel grid used in the calculation is determined according to the point cloud density.

[0058] Specifically, the point spacing, or point density, of the point cloud is first calculated. The initial voxel grid size is set to the point spacing plus 0.01m. After each voxelization process, the current point cloud point spacing is calculated. When the point spacing is less than twice the initial point spacing, the voxel grid size is increased by another 0.01m and the above voxelization process is repeated until the point spacing of the voxelized point cloud is greater than twice the initial point spacing, at which point the voxel grid size no longer increases. The point cloud obtained at this time is point cloud P3.

[0059] Step 3: Calculate the roughness of each point in the voxelized point cloud P3. That is, fit the neighborhood plane of each point using the least squares method and calculate the distance from each point to its neighborhood fitted plane. This distance is the roughness.

[0060] Specifically, for each point, the roughness value is equal to the distance between that point and its nearest best-fit plane. The best-fit plane is calculated using the least squares method, which obtains the unknown parameters. Assuming the plane has three-dimensional data, the plane is defined as follows:

[0061] ax + by + cz + d = 0 (1)

[0062] In the formula, a, b, c, and d are plane parameters;

[0063] For the entire point cloud, construct the covariance matrix:

[0064]

[0065]

[0066] In the formula, p i ∈P; Let be the center point of all points in the point set P.

[0067] According to the formula λV=MV, the eigenvalues ​​and eigenvectors of the covariance M are calculated, where λ=(λ0,λ1,λ2) are the eigenvalues ​​of M, and V=(V0,V1,V2) are the eigenvectors corresponding to λ. By obtaining the minimum value of each eigenvalue, the four parameters a, b, c, and d are calculated to obtain the best-fit plane, and the distance from the point to this plane is calculated as the roughness value.

[0068] Step 4: Assign the calculated roughness value to point cloud P3 to obtain point cloud P4 with roughness features. Extract points whose roughness values ​​meet the conditions as seed points to form seed point cloud P5.

[0069] Specifically, when performing roughness extraction, the roughness value of the point cloud that meets the requirements should be less than 0.02.

[0070] Step 5: Traverse every point in the seed point cloud P5, apply the kdtree-based method to the point cloud P2 for each point in the P5 point cloud, and save all the searched points as a new point cloud P6.

[0071] Specifically, traverse point cloud P5, perform nearest neighbor search on point cloud P5 with a search radius of 0.3, search object is point cloud P2, and save the searched points as a new point cloud P6.

[0072] Step 6: Delete duplicate points in point cloud P6 to obtain secondary lining point cloud P7;

[0073] The kd-tree-based method searches all points in point cloud P6 with a radius of 0.00001. If there are more points within the radius, the more points are deleted. The point cloud after deletion is the secondary lining point cloud P7.

[0074] Step 7: Use the RANSAC algorithm to fit the plane A where the bottom point cloud is located on the secondary lining point cloud P7, and project the point cloud P7 onto the plane A to obtain the planar point cloud P8.

[0075] Specifically, when fitting a plane for point cloud projection, the plane distance threshold is made less than D and the expected probability of outliers is satisfied by iterative loops to ensure that the fitted plane is on the same plane as the original ground.

[0076] Step 8: Calculate the covariance matrix for each point in the planar point cloud P8, and extract the point cloud that satisfies the linear threshold as the linear region point cloud P9;

[0077] Specifically, the method for calculating the covariance of point cloud P8 is to refer to formulas (2) and (3) above. The calculated covariance characteristic value is λ=(λ0,λ1,λ2), where λ0>λ1>λ2;

[0078] The definition of a linear value is:

[0079]

[0080] The linear threshold is 0.9. The point cloud that meets the linear threshold is extracted, that is, all points with a linear value greater than 0.9 are extracted after calculation to form the linear region point cloud P9.

[0081] Step 9: Extract straight lines from the linear region point cloud P9 using multi-line fitting based on the RANSAC method;

[0082] Specifically, the maximum number of iterations for fitting is 1000, the number of loops is 6, and the number of points on each line must be greater than 1000.

[0083] Step 10: Calculate the maximum length of each line point cloud from the extracted multiple line point clouds, and take the average of this length as the distance for the secondary lining advance.

[0084] The original point cloud in this embodiment is as follows: Figure 2 As shown, Figure 3 This is the voxelized point cloud. Figure 4 To meet the roughness requirements of the seed point cloud, Figure 5 This is a two-layered point cloud obtained by secondary segmentation of the seed point cloud. Figure 6 The advance distance obtained from the secondary lining point cloud is... Figure 6The bottom layer is the projected point cloud of the second-to-last point cloud, the rectangular point cloud of the second-to-last layer is the point cloud obtained by extracting linear regions from the projected point cloud, and the two linear point clouds above the rectangular point cloud are the straight line point clouds obtained by fitting.

[0085] Example 2

[0086] This embodiment discloses a system for segmenting point clouds and extracting construction progress information for secondary tunnel lining.

[0087] A point cloud segmentation and construction progress information extraction system for tunnel secondary lining includes:

[0088] The roughness calculation module is configured to: denoise and voxelize the tunnel point cloud, fit a neighborhood plane to each point in the voxelized point cloud and calculate the distance from each point to its neighborhood fitted plane, and calculate the roughness.

[0089] The roughness assignment and filtering module is configured to: assign the calculated roughness value to the voxelized point cloud to obtain a point cloud with roughness features, and extract points whose roughness values ​​meet the set conditions as seed points to form a seed point cloud.

[0090] The radius search module is configured to: traverse every point in the seed point cloud, perform a radius search on the denoised point cloud, and save the searched points as a new point cloud;

[0091] The duplicate point deletion module is configured to delete duplicate points in the new point cloud to obtain a secondary lining point cloud.

[0092] The projection module is configured to: fit the plane A containing the bottom point cloud to the secondary liner point cloud, and project the secondary liner point cloud onto plane A to obtain the planar point cloud;

[0093] The linear region point cloud acquisition module is configured to: calculate the covariance matrix for each point in the planar point cloud and calculate the linear value; extract the point cloud that meets the linear threshold as the linear region point cloud; and extract straight lines from the linear region point cloud.

[0094] The secondary lining advance distance calculation module is configured to: calculate the maximum length of each extracted straight point cloud, and take the average of the maximum lengths as the secondary lining advance distance.

[0095] Example 3

[0096] The purpose of this embodiment is to provide a computer-readable storage medium.

[0097] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the method for segmenting point clouds and extracting construction progress information for secondary tunnel lining as described in Embodiment 1 of this disclosure.

[0098] Example 4

[0099] The purpose of this embodiment is to provide an electronic device.

[0100] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the method for segmenting point clouds and extracting construction progress information for secondary tunnel lining as described in Embodiment 1 of this disclosure.

[0101] The steps and methods involved in the apparatuses of Embodiments 2, 3, and 4 above correspond to those in Embodiment 1. For specific implementation details, please refer to the relevant description section of Embodiment 1. The term "computer-readable storage medium" should be understood as a single medium or multiple media including one or more instruction sets; it should also be understood as including any medium capable of storing, encoding, or carrying an instruction set for execution by a processor and enabling the processor to perform any of the methods in this invention.

[0102] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0103] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for segmenting point clouds and extracting construction progress information for secondary tunnel lining, characterized in that, Includes the following steps: The tunnel point cloud is denoised and voxelized. For each point in the voxelized point cloud, a neighborhood plane is fitted and the distance from each point to its neighborhood fitted plane is calculated. The roughness is then calculated. The calculated roughness values ​​are assigned to the voxelized point cloud to obtain a point cloud with roughness features. Points with roughness values ​​that meet the set conditions are extracted as seed points to form a seed point cloud. Traverse every point in the seed point cloud, perform a radius search on the denoised point cloud, and save the searched points as a new point cloud; Delete duplicate points in the new point cloud to obtain a secondary point cloud; Fit the point cloud of the secondary lining to obtain the plane A containing the bottom point cloud, and project the secondary lining point cloud onto plane A to obtain the planar point cloud; For each point in the planar point cloud, calculate the covariance matrix and the linear value. Extract the point cloud that satisfies the linear threshold as the linear region point cloud, and extract the straight line from the linear region point cloud. For the extracted point clouds of multiple straight lines, calculate the maximum length of each point cloud, and take the average of the maximum lengths as the distance for the secondary lining advance.

2. The method for point cloud segmentation and construction progress information extraction of tunnel secondary lining as described in claim 1, characterized in that: A combination of statistical filtering and radius filtering was used to denoise the tunnel point cloud. When performing voxelization on the denoised point cloud, the size of the voxel grid is determined based on the point cloud density.

3. The method for point cloud segmentation and construction progress information extraction of tunnel secondary lining as described in claim 1, characterized in that, For each point in the voxelized point cloud, the least squares method is used to fit the neighborhood plane.

4. The method for point cloud segmentation and construction progress information extraction of tunnel secondary lining as described in claim 1, characterized in that, Traverse the seed point cloud, perform a nearest neighbor search on the seed point cloud, set the search radius, and search the denoised point cloud. Save the searched points as a new point cloud.

5. The method for point cloud segmentation and construction progress information extraction of tunnel secondary lining as described in claim 1, characterized in that, The kd-tree-based method searches all points in the new point cloud within a set radius. If there are redundant points within the set radius, the redundant points are deleted. The point cloud after deletion is the secondary lining point cloud.

6. The method for point cloud segmentation and construction progress information extraction of tunnel secondary lining as described in claim 1, characterized in that, The RANSAC algorithm is used to fit the plane A of the bottom point cloud to the secondary lining point cloud. Through iterative iteration, the plane distance threshold is made less than D and the expected probability of the outlier value satisfies Ex, so as to ensure that the fitted plane is on the same plane as the original ground.

7. The method for point cloud segmentation and construction progress information extraction of tunnel secondary lining as described in claim 1, characterized in that, Calculate the covariance of each point in the planar point cloud. The eigenvalues ​​of the covariance are λ = (λ0, λ1, λ2), where λ0 > λ1 > λ2. The linear value is defined as: Set a linear threshold, and extract all points whose calculated linear values ​​in the point cloud are greater than the linear threshold to form a linear region point cloud.

8. A system for segmenting point clouds and extracting construction progress information for secondary tunnel lining, characterized in that: include: The roughness calculation module is configured to: denoise and voxelize the tunnel point cloud, fit a neighborhood plane to each point in the voxelized point cloud and calculate the distance from each point to its neighborhood fitted plane, and calculate the roughness. The roughness assignment and filtering module is configured to: assign the calculated roughness value to the voxelized point cloud to obtain a point cloud with roughness features, and extract points whose roughness values ​​meet the set conditions as seed points to form a seed point cloud. The radius search module is configured to: traverse every point in the seed point cloud, perform a radius search on the denoised point cloud, and save the searched points as a new point cloud; The duplicate point deletion module is configured to delete duplicate points in the new point cloud to obtain a secondary lining point cloud. The projection module is configured to: fit the plane A containing the bottom point cloud to the secondary liner point cloud, and project the secondary liner point cloud onto plane A to obtain the planar point cloud; The linear region point cloud acquisition module is configured to: calculate the covariance matrix for each point in the planar point cloud and calculate the linear value; extract the point cloud that meets the linear threshold as the linear region point cloud; and extract straight lines from the linear region point cloud. The secondary lining advance distance calculation module is configured to: calculate the maximum length of each extracted straight point cloud, and take the average of the maximum lengths as the secondary lining advance distance.

9. A computer-readable storage medium having a program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the method for segmenting point clouds and extracting construction progress information for tunnel secondary lining as described in any one of claims 1-7.

10. An electronic device, comprising a memory, a processor, and a program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the method for segmenting point clouds and extracting construction progress information for tunnel secondary lining as described in any one of claims 1-7.

Images