Tunnel segment ring ovality analysis method based on point cloud data and computer device

Abstract

The invention relates to the field of ovality analysis, in particular to a tunnel segment ring ovality analysis method based on point cloud data and a computer device. According to the scheme, a plurality of threads are created, each thread is responsible for reading a part of data, streaming reading is adopted, after a certain scale of point data is obtained every time, the data are integrated in a main storage, and the read point cloud data are preprocessed through an octree, including dividing the point cloud data and storing the divided data in a cache; quickly positioning and extracting required data by using an index mechanism of an octree; the data of all the leaf nodes meeting the query conditions are merged, and a final query result is generated; according to the final query result, calculating the mileage, the horizontal offset distance and the vertical offset distance of each point through a preset alignment algorithm, and designing contour difference value information; and extracting a to-be-analyzed section and forming a slice packet, then performing optimal ellipse search, and finally calculating ellipse parameters and returning ellipticity. The method is suitable for ellipticity analysis.

Description

Technical Field

[0001] The invention relates to the field of ellipticity analysis, and in particular to a tunnel segment ring ellipticity analysis method based on point cloud data and a computer device. Background Art

[0002] In tunnel engineering, the geometry of the segment ring is crucial to the stability and durability of the tunnel structure. Theoretically, shield tunnel segments are required to be circular, but since shield segments may deviate during production, installation, and use, ellipticity is used to quantify the geometry of shield tunnel segments. Ellipticity is an important parameter for quantitatively analyzing the overall deformation of shield segments and is an important link in ensuring the stability and safety of tunnel structures. The relevant specifications for the design, acceptance, and protection of the shield method for tunneling have regulations and requirements for the ellipticity detection of shield tunnel segments.

[0003] Traditional segment ring inspection methods mainly rely on manual measurement and two-dimensional image analysis. Surveyors use total stations to measure selected points in the ring at designated cross-section locations. There are usually two methods: direct method and indirect method: Direct method: Directly measure and determine the major and minor axes of the shield segment ellipse. The most direct measurement method is the plumb line method. This method uses a tape measure and a plumb line to measure the major and minor axes of the tunnel segment. The major axis is measured multiple times in the horizontal direction with a tape measure, and the longest side is considered to be the major axis; the minor axis can be determined by hanging a plumb line so that the plumb line passes through the center of the major axis. The premise for the application of this method is that the minor semi-axis of the ellipse is parallel to the direction of the plumb line, but the actual situation is that the major semi-axis of the tunnel is not necessarily in the horizontal direction, so it has a certain narrowness.

[0004] Indirect method: The indirect method does not directly measure the major and minor axes. Instead, it measures the tunnel surface profile and uses ellipse fitting to obtain the major and minor axes of the ellipse and the rotation angle of the ellipse. Compared with the direct method, it has a high degree of automation and more accurate measurement results.

[0005] The above traditional methods can realize the ovality detection of shield tunnel segments to a certain extent, but they have the following main defects and shortcomings: The measurement error is large and it is difficult to ensure accuracy: The manual measurement method relies on the operating level of the measurement personnel and is easily affected by human factors, resulting in large measurement errors and difficulty in ensuring measurement accuracy.

[0006] Data processing is complex and inefficient: Traditional methods require manual recording and organization of data. The data processing process is cumbersome and error-prone, making it difficult to achieve efficient analysis and processing.

[0007] Unable to fully reflect the three-dimensional shape of the segment ring: Traditional methods are mainly based on two-dimensional plane measurement, which makes it difficult to fully reflect the three-dimensional geometric shape of the segment ring and cannot fully understand its complex spatial deformation characteristics. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a tunnel segment ring ellipticity analysis method and computer device based on point cloud data, thereby improving measurement accuracy and data processing efficiency and achieving more accurate ellipticity analysis.

[0009] The present invention adopts the following technical solutions to achieve the above-mentioned purpose. In a first aspect, the present invention provides a tunnel segment ring ellipticity analysis method based on point cloud data, comprising: S1, read the point cloud data of the tunnel scan; S2, preprocessing the read point cloud data through the octree; S201, dividing the point cloud data; S202, storing the divided data in a cache; S203, using the octree indexing mechanism to quickly locate and query required data; S204, merging all data that meet the query conditions to generate a query result; S3. Calculate the mileage, horizontal offset, vertical offset and design profile information of each point cloud data by a preset alignment algorithm, wherein the point cloud data is the point cloud data read from the tunnel scan or the point cloud data in the query result; S4, extracting the section to be analyzed and forming a slice package; The user can directly provide a slice list of the section to be analyzed according to their own needs, or provide relevant parameters to automatically generate a slice list. According to the slice list and the thickness set by the preset parameters, the start and end mileage range of each slice is calculated, and the corresponding point cloud data is selected from the pre-processed point cloud data according to the start and end mileage range as the point cloud data of the slice to be analyzed; S5, search for the best ellipse based on RANSAC; Perform geometric analysis on the slices in the cross-section slice package according to the horizontal offset and vertical offset data, and remove invalid cross-section slices; The original slice data is preliminarily processed based on the standard circle, and noise points are eliminated according to the difference of the designed contour; Perform a fast least squares fit on the preliminarily processed data to find the preliminary best fitting ellipse; Calculate the score of the preliminary fitted ellipse and confirm the set of interior points; Based on the internal point set, multiple points are randomly obtained, the ellipse is refitted, and the new fitting ellipse score is calculated. If the score increases, the previous score result is replaced until the score is stable or the number of cycles reaches the upper limit; S6, calculate the ellipse parameters and return the ellipse degree; According to the final fitting ellipse equation, its major axis and minor axis parameters are calculated, and the difference between the major axis and the minor axis is divided by the diameter of the tunnel to obtain the ellipticity.

[0010] Furthermore, step S1 specifically includes: Create multiple threads, each thread is responsible for reading a part of the data, using streaming reading. After obtaining the data of the set scale point, the data is integrated into the main storage.

[0011] Furthermore, step S201 specifically includes: Determine the spatial extent of the entire point cloud data; Initialize the root node, which represents the entire spatial range and serves as the starting point of the octree; Starting from the root node, the space is divided into eight subspaces according to the distribution of point cloud data. Each subspace corresponds to a subnode, representing eight equally divided areas of the space. Assign point cloud data to corresponding child nodes. If the point cloud data contained in a child node exceeds the set threshold, continue to divide the child node until the preset conditions are met. When the number of point cloud data in the child node is lower than the preset threshold, or the spatial size of the child node is smaller than the set minimum unit, stop dividing.

[0012] Furthermore, step S202 specifically includes: Each octree node contains the spatial boundary information of the node, point cloud data, and pointers to its child nodes; The divided point cloud data is stored in the cache file according to the nodes; A unique identifier is generated for each node to record and retrieve its corresponding point cloud data, and a file path index is generated according to the node's level and position.

[0013] Furthermore, step S203 specifically includes: Locate the nodes within the target space through the nodes of the octree; Using the spatial boundary information of the node, starting from the root node, the query is performed layer by layer according to the spatial scope of the query. The query process specifically includes: Starting from the root node, based on the query's spatial range, determine whether the query range overlaps with the spatial range of the current node. If so, perform a recursive check until all leaf nodes that overlap with the query range are found; After the query is completed, data extraction is performed, including: According to the located leaf node, the corresponding point cloud data is extracted from the cache. If the cache data is stored in the memory, the point cloud data of the corresponding node is directly read. If the cache data is stored in the disk, the corresponding file or database record is read according to the path.

[0014] Furthermore, in step S3, calculating the mileage, horizontal offset, vertical offset and design contour difference information of each point cloud data specifically includes: Determine the horizontal curve segment to which any point P belongs: Given the coordinates of point P (X p , Y P , Z P ), determine on which horizontal curve alignment line element point P is located through the spatial distance relationship between point P and the endpoints of each line element; Project point P vertically onto the horizontal curve alignment element where it is located, and obtain the coordinates of the projection point P'; The length L from the starting point of the horizontal curve to the projection point P', plus the initial mileage K of the starting point of the horizontal curve 0 , we can get the mileage of point P as K = L + K 0 ; Calculate the vertical distance from point P to its projection point P' to obtain the horizontal offset O; According to the mileage K of point P, combined with the vertical curve definition, calculate the design elevation data H corresponding to the mileage K 0 ; The vertical offset H of point P is equal to the Z coordinate value Z of point P p Subtract the design elevation H 0 ; The design contour difference D is equal to the distance from point P to the design center minus the design radius.

[0015] Furthermore, removing noise points according to the design outline specifically includes: If the distance between the projection of the point cloud in the original slice on the design contour and the point cloud exceeds the set threshold, it is judged as a noise point and is removed.

[0016] In a second aspect, the present invention provides a computer device, including a memory, wherein the memory stores program instructions, and when the program instructions are run, the above-mentioned tunnel segment ring ellipticity analysis method based on point cloud data is executed.

[0017] The beneficial effects of the present invention are: The present invention performs analysis based on point cloud data, uses an octree to preprocess the point cloud data, and spatially divides the point cloud data through an octree data structure, thereby achieving efficient classification, cutting and caching of the point cloud data.

[0018] When searching for the best ellipse based on the RANSAC concept, the present invention performs validity checks on the cross-section slices, removes invalid cross-sections, performs preliminary filtering, filters out obvious noise points, and combines the advantages of the least squares method and the random sampling consistency algorithm. Users can choose a fast least squares fitting method according to actual conditions, or combine the characteristics of the RANSAC algorithm that is more sensitive to local features, and select a fitting ellipse that is more in line with the actual situation through a scoring mechanism. This diversified fitting strategy not only improves the fitting accuracy, but also enhances the robustness of the algorithm, and ultimately achieves a more accurate ellipticity analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a flow chart of a tunnel segment ring ellipticity analysis method based on point cloud data provided by an embodiment of the present invention; Figure 2 It is a flow chart of calculating service information by a route determination algorithm provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0020] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0021] The present invention provides a tunnel segment ring ellipticity analysis method based on point cloud data, such as Figure 1 As shown, including: S1, read the point cloud data of the tunnel scan; Point cloud data is read simultaneously through multi-threading technology, which is implemented as follows: Create multiple threads, each thread is responsible for reading a part of the data; Using streaming reading, after obtaining a certain scale of point data, the data is integrated in the main storage and prepared for further processing.

[0022] S2, preprocessing the read point cloud data through the octree; Construct an octree data structure: Define the spatial range and determine the spatial range of the entire point cloud data; Initialize the root node, which represents the entire space range and is the starting point of the octree; Recursively partition the space: Starting from the root node, the space is divided into eight subspaces according to the distribution of point cloud data, which are the eight subnodes of the octree.

[0023] Each subspace corresponds to a child node, representing eight equally divided regions of the space.

[0024] Distribute point cloud data: The point cloud data is assigned to the corresponding sub-nodes. If the point cloud data contained in a sub-node exceeds a certain threshold, the sub-node will continue to be further divided until the preset conditions are met.

[0025] When the amount of point cloud data in a child node is lower than a preset threshold, or the spatial size of the child node is smaller than a certain minimum unit, the division is stopped.

[0026] The divided data is stored in the cache, including: Node data structure design, each octree node contains the spatial boundary information of the node, the point cloud data it contains, and a pointer to its child nodes.

[0027] The cache strategy stores the divided point cloud data in the cache file according to the nodes.

[0028] Storage path design: A unique identifier is generated for each node, which is used to record and retrieve its corresponding point cloud data.

[0029] A file path index can be generated based on the node's level and position.

[0030] Utilize the efficient indexing mechanism of the octree to quickly locate and extract the required data: Index mechanism design: The node structure of the octree allows for fast location of nodes within the target space.

[0031] Using the spatial boundary information of the node, starting from the root node, the node is positioned layer by layer according to the spatial scope of the query.

[0032] Query process: Starting from the root node, based on the spatial scope of the query, determine whether the query scope overlaps with the spatial scope of the current node. If so, further check the child nodes of the node, using recursive checking until all leaf nodes that overlap with the query scope are found.

[0033] Data extraction: According to the located leaf node, the corresponding point cloud data is extracted from the cache. If the cache data is stored in the memory, the point cloud data of the corresponding node is directly read. If the cache data is stored in the disk, the corresponding file or database record is read according to the path.

[0034] Result merging: Merge the data of all leaf nodes that meet the query conditions to generate the final query result.

[0035] S3, route determination algorithm calculates business information; Based on the preset alignment algorithm, the point cloud in the current main storage or the queried point cloud is calculated in real time, and the business information of the point cloud is stored. This process includes computing the mileage, horizontal offset, vertical offset, design contour difference and other information for each point cloud. After the business information is attached, the data is cached to the file system or other low-speed large-capacity storage system and then the next batch of data is read in a streaming manner.

[0036] like Figure 2 As shown, the specific calculation steps are as follows: Determine the horizontal curve segment to which any point P belongs. Given the coordinates of point P (X p , Y P , Z P ), through the spatial distance relationship between point P and the end points of each line element, determine on which section of the horizontal curve alignment line element the point is located; Project point P vertically onto the horizontal curve alignment element where it is located, and obtain the coordinates of the projection point P'; Calculate the mileage K of point P, the length L from the starting point of the horizontal curve to the projection point P', plus the initial mileage K of the starting point of the horizontal curve 0 , we can get the mileage of point P as K = L + K 0 ; Calculate the horizontal offset O of point P, calculate the vertical distance from point P to its projection point P', and get the offset O; According to the mileage K of point P, combined with the vertical curve definition, calculate the design elevation data H corresponding to the mileage K 0 ; The vertical offset H of point P is equal to the Z coordinate value Z of point P p Subtract the design elevation H 0 ; The design contour difference D is equal to the distance from point P to the design center minus the design radius. The distance from point P to the design center is calculated based on the square sum of O / H.

[0037] S4, quickly extract the section to be analyzed and form a slice package; According to the preset detection requirements, the corresponding sections are extracted, and then the extracted section data are packaged into slice packages to reduce the amount of data for a single calculation.

[0038] The specific implementation steps are as follows: The user can directly provide a slice list of the section to be analyzed according to his own needs, or provide necessary parameters for the system to automatically generate a slice list, such as providing the start and end positions and the spacing between slices; The system calculates the start and end mileage range of each slice according to the specified mileage list and the thickness set by the preset parameters; By utilizing the parallel processing mechanism, point clouds are selected from the pre-processed point clouds according to the start and end mileage range as slices to be analyzed.

[0039] S5, search for the best ellipse based on RANSAC; Validity check: Perform geometric analysis on the slices in the cross-section slice package based on the horizontal offset and vertical offset data, and eliminate invalid cross-section slices.

[0040] Preliminary filtering: Preliminary processing of the original slice data based on the standard circle, and elimination of noise points according to the design contour difference, including: If the distance between the projection of the point cloud in the original slice on the design contour and the point cloud exceeds the set threshold, it is judged as a noise point and is removed.

[0041] Least squares fitting: Perform a fast least squares fitting on the data after preliminary processing to find the preliminary optimal fitting ellipse.

[0042] Score calculation and confirmation of the internal point set: Calculate the score of the preliminary fitted ellipse and confirm the internal point set.

[0043] Refitting and loop calculation: Randomly obtain multiple points based on the inner point set, refit the ellipse, and calculate the new fitted ellipse score. If the score increases, it will replace the previous result until the score stabilizes or the number of loops reaches the upper limit.

[0044] S6, calculate the ellipse parameters and return the ellipse degree; According to the final fitting ellipse equation, its major axis, minor axis and other parameters are calculated, and the difference between the major axis and the minor axis is divided by the diameter of the tunnel to obtain the ellipticity. Finally, the ellipticity is returned to measure the quality of the fitting result.

[0045] By separating the need for report analysis from the actual conflict of the large size of the complete point cloud data, the present invention can maximize the advantages of parallel computing. Specific implementation methods include data sharding, local processing, and integrated analysis. Data sharding technology divides large-scale point clouds into several small blocks that can be processed in parallel, and each small block can perform noise processing and feature extraction independently. The local processing results are quickly integrated through parallel computing to finally generate a complete analysis report. This method not only improves computing efficiency, but also solves the problems of memory overflow and processing bottlenecks caused by excessive data.

[0046] In order to balance the denoising cost and the subsequent computational cost, the present invention provides a method for quickly and preliminarily processing point cloud noise. This method includes multi-level filtering and an adaptive threshold strategy. Primary filtering is used to remove significant noise, and subsequent refinement is performed based on the characteristics of the noise distribution. The overall processing flow minimizes the computational overhead in the denoising process while ensuring basic clarity. In addition, the method can flexibly adjust the denoising intensity to adapt to application scenarios with different accuracy requirements.

[0047] Compared with the traditional least squares fitting, the present invention combines the advantages of the least squares method and the random sampling consensus algorithm (RANSAC). Users can choose a fast least squares fitting method according to the actual situation, or combine the characteristics of the RANSAC algorithm that is more sensitive to local features to select a fitting ellipse that is more in line with the actual situation through a scoring mechanism. This diversified fitting strategy not only improves the fitting accuracy, but also enhances the robustness of the algorithm, especially when there are a large number of outliers.

[0048] In engineering practice, the fitting evolution process of point cloud data is controllable and traceable, which provides users with a basis for multi-party data verification when adopting new methods. The present invention can provide users with intuitive process backtracking and theoretical analysis tools by recording each step of the fitting process and parameter changes in detail. This not only helps users understand and adjust fitting parameters, but also enhances the regulatory transparency in engineering measurement, which is conducive to the promotion and application of new methods. Specifically, users can view the intermediate results and error distribution in the fitting process through visualization tools, so as to better understand data differences and optimization methods.

[0049] In summary, the present invention can more accurately and comprehensively reflect the geometric shape of the tunnel segment rings and improve the stability and safety of the tunnel structure.

[0050] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A tunnel segment ring ellipticity analysis method based on point cloud data, characterized in that: include: S1, read the point cloud data of the tunnel scan; S2, preprocessing the read point cloud data through the octree; S201, dividing the point cloud data; S202, storing the divided data in a cache; S203, using the octree indexing mechanism to quickly locate and query required data; S204, merging all data that meet the query conditions to generate a query result; S3. Calculate the mileage, horizontal offset, vertical offset and design profile information of each point cloud data by a preset alignment algorithm, wherein the point cloud data is the point cloud data read from the tunnel scan or the point cloud data in the query result; S4, extracting the section to be analyzed and forming a slice package; The user can directly provide a slice list of the section to be analyzed according to their own needs, or provide relevant parameters to automatically generate a slice list. According to the slice list and the thickness set by the preset parameters, the start and end mileage range of each slice is calculated, and the corresponding point cloud data is selected from the pre-processed point cloud data according to the start and end mileage range as the point cloud data of the slice to be analyzed; S5, search for the best ellipse based on RANSAC; Perform geometric analysis on the slices in the cross-section slice package according to the horizontal offset and vertical offset data, and remove invalid cross-section slices; The original slice data is preliminarily processed based on the standard circle, and noise points are eliminated according to the difference of the designed contour; Perform a fast least squares fit on the preliminarily processed data to find the preliminary best fitting ellipse; Calculate the score of the preliminary fitted ellipse and confirm the set of interior points; Based on the internal point set, multiple points are randomly obtained, the ellipse is refitted, and the new fitting ellipse score is calculated. If the score increases, the previous score result is replaced until the score is stable or the number of cycles reaches the upper limit; S6, calculate the ellipse parameters and return the ellipse degree; According to the final fitting ellipse equation, its major axis and minor axis parameters are calculated, and the difference between the major axis and the minor axis is divided by the diameter of the tunnel to obtain the ellipticity.

2. The method for analyzing the ellipticity of a tunnel segment ring based on point cloud data according to claim 1 is characterized in that: Step S1 specifically includes: Create multiple threads, each thread is responsible for reading a part of the data, using streaming reading. After obtaining the data of the set scale point, the data is integrated into the main storage.

3. The method for analyzing the ellipticity of a tunnel segment ring based on point cloud data according to claim 1 is characterized in that: Step S201 specifically includes: Determine the spatial extent of the entire point cloud data; Initialize the root node, which represents the entire spatial range and serves as the starting point of the octree; Starting from the root node, the space is divided into eight subspaces according to the distribution of point cloud data. Each subspace corresponds to a subnode, representing eight equally divided areas of the space. Assign point cloud data to corresponding child nodes. If the point cloud data contained in a child node exceeds the set threshold, continue to divide the child node until the preset conditions are met. When the number of point cloud data in the child node is lower than the preset threshold, or the spatial size of the child node is smaller than the set minimum unit, stop dividing.

4. The method for analyzing the ellipticity of a tunnel segment ring based on point cloud data according to claim 3 is characterized in that: Step S202 specifically includes: Each octree node contains the spatial boundary information of the node, point cloud data, and pointers to its child nodes; The divided point cloud data is stored in the cache file according to the nodes; A unique identifier is generated for each node to record and retrieve its corresponding point cloud data, and a file path index is generated according to the node's level and position.

5. The method for analyzing the ellipticity of a tunnel segment ring based on point cloud data according to claim 4 is characterized in that: Step S203 specifically includes: Locate the nodes within the target space through the nodes of the octree; Using the spatial boundary information of the node, starting from the root node, the query is performed layer by layer according to the spatial scope of the query. The query process specifically includes: Starting from the root node, based on the query's spatial range, determine whether the query range overlaps with the spatial range of the current node. If so, perform a recursive check until all leaf nodes that overlap with the query range are found; After the query is completed, data extraction is performed, including: According to the located leaf node, the corresponding point cloud data is extracted from the cache. If the cache data is stored in the memory, the point cloud data of the corresponding node is directly read. If the cache data is stored in the disk, the corresponding file or database record is read according to the path.

6. The method for analyzing the ellipticity of a tunnel segment ring based on point cloud data according to claim 1, characterized in that: In step S3, calculating the mileage, horizontal offset, vertical offset and design contour difference information of each point cloud data specifically includes: Determine the horizontal curve segment to which any point P belongs: Given the coordinates of point P (X p , Y P , Z P ), determine on which horizontal curve alignment line element point P is located through the spatial distance relationship between point P and the endpoints of each line element; Project point P vertically onto the horizontal curve alignment element where it is located, and obtain the coordinates of the projection point P'; The length L from the starting point of the horizontal curve to the projection point P', plus the initial mileage K0 of the starting point of the horizontal curve, gives the mileage of point P K = L + K0; Calculate the vertical distance from point P to its projection point P' to obtain the horizontal offset O; According to the mileage K of point P and the vertical curve definition, calculate the design elevation data H0 corresponding to the mileage K; The vertical offset H of point P is equal to the Z coordinate value Z of point P p Subtract the design elevation H0; The design contour difference D is equal to the distance from point P to the design center minus the design radius.

7. The method for analyzing the ellipticity of a tunnel segment ring based on point cloud data according to claim 1, characterized in that: Eliminating noise points according to the design outline specifically includes: If the distance between the projection of the point cloud in the original slice on the design contour and the point cloud exceeds the set threshold, it is judged as a noise point and is removed.

8. A computer device, comprising a memory, wherein the memory stores program instructions, characterized in that: When the program instructions are executed, the tunnel segment ring ellipticity analysis method based on point cloud data as described in any one of claims 1 to 7 is executed.

Images