A tunnel structural plane identification method, device and equipment and a storage medium

Abstract

The application discloses a tunnel structural plane identification method, device and equipment and a storage medium, and is applied to the technical field of tunnel engineering geology analysis, and comprises the following steps: acquiring all three-dimensional point cloud data of a tunnel structural plane; performing main body extension direction analysis on all the three-dimensional point cloud data to obtain slice direction information; performing slice processing on all the three-dimensional point cloud data based on the slice direction information to obtain two-dimensional tunnel slice data; performing feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features; performing spatial deviation field processing on the spatial distribution features based on a pre-constructed polar radius constraint model to obtain a spatial deviation result; performing feature screening on all the three-dimensional point cloud data based on the spatial deviation result to obtain to-be-fitted point cloud data; performing two-stage fitting on all the to-be-fitted point cloud data, and determining the spatial orientation of the tunnel structural plane based on a two-stage fitting result. The method can accurately identify the spatial orientation of the tunnel structural plane.

Description

Technical Field

[0001] This invention relates to the field of geological analysis technology for tunnel engineering, and in particular to a method, apparatus, equipment and storage medium for identifying tunnel structural surfaces. Background Technology

[0002] The spatial orientation of the tunnel structure is a core geological parameter for assessing the stability of the surrounding rock, designing support systems, and predicting potential landslide risks. The reliability of its measurement results directly affects the safety and economy of engineering decisions.

[0003] In existing technologies, tunnel structural surface identification methods rely on geological technicians using handheld instruments such as geological compasses to directly measure the spatial orientation of the tunnel structural surfaces on the exposed rock surfaces. However, this method is easily affected by the surveyor's experience, perspective bias, and environmental conditions such as lighting, temperature, and humidity within the tunnel, leading to significant measurement errors. Furthermore, blind spots and safety risks exist within the tunnel, making it impossible to obtain complete tunnel structural surface data. This results in incomplete geological models, which are insufficient to meet the requirements for safe construction and long-term stable operation of tunnel projects. Summary of the Invention

[0004] This invention provides a method, apparatus, device, and storage medium for identifying tunnel structural surfaces, in order to solve the technical problem of large errors in the results of manual measurement of the spatial orientation of tunnel structural surfaces, and to achieve accurate identification of the spatial orientation of tunnel structural surfaces.

[0005] To address the aforementioned technical problems, this invention provides a method for identifying tunnel structural surfaces, comprising: Acquire all 3D point cloud data of the tunnel structure surface; Perform main body extension direction analysis on all the aforementioned 3D point cloud data to obtain slice direction information; Based on the slice direction information, slice processing is performed on all the three-dimensional point cloud data to obtain two-dimensional tunnel slice data; Feature extraction processing is performed on all the two-dimensional tunnel slice data to obtain spatial distribution features; Based on the pre-constructed extreme diameter constraint model, the spatial distribution characteristics are processed by spatial deviation field to obtain spatial deviation results. The extreme diameter constraint model is constructed from the analysis results of the actual geometry of the tunnel structure surface, and the spatial deviation field processing is configured to analyze the degree of theoretical and actual deviation of the spatial distribution characteristics of the tunnel structure surface. Based on the spatial deviation results, feature filtering is performed on all the three-dimensional point cloud data to obtain the point cloud data to be fitted. A two-stage fitting is performed on all the point cloud data to be fitted, and the spatial orientation of the tunnel structure surface is determined based on the two-stage fitting results.

[0006] As one preferred embodiment, the step of performing main extension direction analysis on all the three-dimensional point cloud data to obtain slice direction information includes: The spatial location information corresponding to each extracted 3D point cloud data is analyzed to obtain the spatial location covariance matrix. Based on the principal component analysis method, the spatial location covariance matrix is ​​subjected to eigenvalue decomposition to obtain the principal eigenvectors; The slice direction information is determined based on the main feature vector.

[0007] As one preferred embodiment, the step of slicing all the three-dimensional point cloud data based on the slicing direction information to obtain two-dimensional tunnel slice data includes: Using the axial direction indicated by the slice direction information as a reference, all the three-dimensional point cloud data are sliced ​​at equal intervals to obtain point cloud slice data, wherein the interval of the equal interval slice processing is greater than the average density of the three-dimensional point cloud data. Based on the axis direction, a two-dimensional projection transformation is performed on all the point cloud slice data to obtain the corresponding two-dimensional tunnel slice data.

[0008] As one preferred embodiment, the step of performing feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features includes: Extract the spatial coordinate information of all the two-dimensional tunnel slice data; The spatial coordinate information is transformed into polar coordinates to obtain the polar coordinate information corresponding to each of the two-dimensional tunnel slice data. Based on the polar coordinate information, feature extraction processing is performed on the corresponding two-dimensional tunnel slice data to obtain the spatial distribution features.

[0009] As one preferred embodiment, the spatial distribution characteristics are processed using a spatial deviation field based on a pre-constructed polar radius constraint model to obtain spatial deviation results, including: Based on the aforementioned extreme diameter constraint model, spatial characteristic analysis is performed on the tunnel structural surface to obtain theoretical spatial distribution characteristics; The spatial distribution characteristics and the theoretical spatial distribution characteristics are subjected to the spatial deviation field processing to obtain the spatial deviation result.

[0010] As one preferred embodiment, the step of performing feature filtering on all the three-dimensional point cloud data based on the spatial deviation results to obtain the point cloud data to be fitted includes: Each of the three-dimensional point cloud data is subjected to feature labeling processing to obtain feature labeling results; Based on the spatial deviation results, all the feature labeling results are filtered to obtain the point cloud data to be fitted.

[0011] As one preferred embodiment, the step of performing a two-stage fitting on all the point cloud data to be fitted, and determining the spatial orientation of the tunnel structural surface based on the two-stage fitting results, includes: Based on a random sampling algorithm, the point cloud data to be fitted is subjected to a first-stage fitting process to obtain an initial fitting result. Based on the least squares method, the initial fitting result is subjected to a second-stage fitting process to obtain the point cloud fitting result; Spatial orientation analysis is performed on the point cloud fitting results to obtain the dip and tilt angle data of the tunnel structure surface; By integrating the dip data and the tilt angle data, the spatial orientation of the tunnel structure surface is obtained.

[0012] Another aspect of the present invention provides a tunnel structure surface identification device, comprising: The data acquisition module is used to acquire all three-dimensional point cloud data of the tunnel structure surface; The orientation analysis module is used to perform main extension orientation analysis on all the three-dimensional point cloud data to obtain slice orientation information; The data slicing module is used to slice all the three-dimensional point cloud data based on the slicing direction information to obtain two-dimensional tunnel slice data; The feature extraction module is used to perform feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features; The deviation processing module is used to perform spatial deviation field processing on the spatial distribution characteristics based on a pre-constructed extreme diameter constraint model to obtain spatial deviation results. The extreme diameter constraint model is constructed from the analysis results of the actual geometry of the tunnel structure surface, and the spatial deviation field processing process is configured to analyze the degree of theoretical and actual deviation of the spatial distribution characteristics of the tunnel structure surface. The data filtering module is used to perform feature filtering on all the three-dimensional point cloud data based on the spatial deviation results to obtain the point cloud data to be fitted. The spatial orientation determination module is used to perform two-stage fitting on all the point cloud data to be fitted, and to determine the spatial orientation of the tunnel structure surface based on the two-stage fitting results.

[0013] In another aspect, the present invention provides a tunnel structure surface identification device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement a tunnel structure surface identification method as described above.

[0014] In another aspect, the present invention provides a computer-readable storage medium storing a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements a tunnel structure surface identification method as described above.

[0015] Compared with existing technologies, the beneficial effects of the embodiments of the present invention are at least one of the following: The present invention achieves non-contact, high-density, and digital acquisition of tunnel surrounding rock by obtaining three-dimensional point cloud data of tunnel structural surfaces, solving the problems of low efficiency and incomplete coverage of traditional manual measurement; The present invention achieves automatic extraction and geometric alignment of tunnel axis direction by analyzing the main extension direction of three-dimensional point cloud data and determining the slicing direction, solving the problem of relying on manual specification or design of the axis and being unable to adapt to changes in actual tunnel morphology; The present invention achieves structured expression of point cloud geometric information by converting three-dimensional point cloud data into two-dimensional slice dataset and extracting features from it, providing a unified comparison benchmark for subsequent deviation analysis; The present invention performs spatial deviation field processing on spatial distribution features based on the polar radius constraint model, realizing a quantitative assessment of the degree of geometric deviation between each point and the design contour; The present invention performs feature screening based on spatial deviation results and performs two-stage fitting on the screened results, realizing automatic identification and removal of non-rock mass interference points, solving the problem that single fitting methods are easily affected by noise, and providing a reliable basis for obtaining accurate spatial orientation of tunnel structural surfaces. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for identifying tunnel structure surfaces in one embodiment of the present invention; Figure 2 This is a diagram showing the preprocessing result of point cloud data in one embodiment of the present invention; Figure 3 This is a schematic diagram of the fitting plane of tunnel surrounding rock joints in one embodiment of the present invention; Figure 4 This is a schematic diagram of the spatial orientation of tunnel surrounding rock joints in one embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of a tunnel structure surface identification device in one embodiment of the present invention; Figure 6 This is a structural block diagram of a tunnel structure surface recognition device according to one embodiment of the present invention; Figure label: The module includes: 11. Data acquisition module; 12. Direction analysis module; 13. Data slicing module; 14. Feature extraction module; 15. Deviation processing module; 16. Data filtering module; and 17. Spatial orientation determination module. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] In the description of this application, the terms "first," "second," "third," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined with "first," "second," "third," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0019] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and do not indicate or imply that the device or component referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0020] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0021] The spatial orientation of tunnel structural faces is a key quantitative indicator in engineering geological analysis and rock mechanics evaluation, playing a decisive role in assessing surrounding rock stability, formulating targeted support schemes, and providing early warning of potential landslide risks. This parameter directly affects the anisotropy of rock mass strength, the location of potential slip boundaries, and the determination of failure modes; if the measurement results are inaccurate, it may lead to the failure of support measures, a surge in engineering costs, or safety accidents. Therefore, obtaining high-precision and reliable spatial orientation of structural faces is a fundamental prerequisite for ensuring the safety and controllability of tunnel engineering, the economic rationality of design, and the long-term stable operation and maintenance.

[0022] Currently, the mainstream method in this field still relies on traditional manual field measurement techniques, which involve geological technicians using handheld instruments such as geological compasses to directly measure the rock surface exposed by the tunnel in order to obtain the attitude data of the structural plane. However, this method has a series of inherent limitations: First, the measurement results are significantly affected by human factors. Different personnel may introduce subjective biases due to differences in experience, operating habits, and judgment standards. Even the same person may experience data consistency issues due to fatigue, limited viewing angle, or reading errors. Second, the working environment is subject to harsh constraints. Insufficient lighting, high humidity, and dust in the tunnel can all interfere with instrument readings and reduce measurement reliability. Third, there are blind spots and safety hazards. Effective measurements are often impossible in areas such as the arch and high sidewalls where personnel cannot easily reach or stay for extended periods, resulting in incomplete geological logging. Furthermore, personnel working in unsupported or unstable tunnel sections face higher safety risks. Fourth, it is inefficient and lacks scalability. In large-section, long-distance tunnel projects, manual point-by-point measurements are time-consuming and laborious, making it difficult to achieve rapid and comprehensive data acquisition and dynamic updates. This fails to meet the urgent need for high-precision, full-section geological information real-time feedback in modern, digital construction and operation.

[0023] One embodiment of the present invention provides a method for identifying tunnel structural surfaces. For details, please refer to [link to relevant documentation]. Figure 1 , Figure 1 The diagram shown is a flowchart of a tunnel structure surface identification method according to one embodiment of the present invention. The method includes steps S1 to S7: S1. Obtain all three-dimensional point cloud data of the tunnel structure surface; S2. Perform main body extension direction analysis on all the three-dimensional point cloud data to obtain slice direction information; S3. Based on the slicing direction information, slice all the three-dimensional point cloud data to obtain two-dimensional tunnel slice data; S4. Perform feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features; S5. Based on the pre-constructed extreme diameter constraint model, the spatial distribution characteristics are processed by spatial deviation field to obtain spatial deviation results. The extreme diameter constraint model is constructed from the analysis results of the actual geometric shape of the tunnel structure surface. The spatial deviation field processing is configured to analyze the degree of theoretical and actual deviation of the spatial distribution characteristics of the tunnel structure surface. S6. Based on the spatial deviation results, perform feature filtering on all the three-dimensional point cloud data to obtain the point cloud data to be fitted. S7. Perform two-stage fitting on all the point cloud data to be fitted, and determine the spatial orientation of the tunnel structure surface based on the two-stage fitting results.

[0024] Traditional manual handheld instrument measurements are susceptible to the influence of personnel experience, operating habits, and environmental factors such as lighting and dust within the tunnel, resulting in large measurement errors and difficulty in covering areas inaccessible to personnel, such as the arch and high sidewalls. This poses safety hazards and is inefficient, failing to meet the engineering requirements for high-precision and comprehensive spatial orientation data of structural surfaces. Therefore, in step S1, this invention further employs a multi-station scanning and stitching mode. Multiple scanning stations are fixed at appropriate heights from the bottom plate along the tunnel axis. The scanner resolution, scanning distance range, and point cloud density are optimized based on the characteristics of the surrounding rock surface. After controlling the scanning time of each station, the multi-station point cloud stitching is completed using the scanner's software in a target-sphere-based feature matching manner, with strict control over stitching accuracy. This ultimately yields the original point cloud data, i.e., the three-dimensional point cloud data. After stitching the three-dimensional point cloud data, denoising is required. Specifically, wavelet transform is used to denoise the three-dimensional point cloud data. For detailed denoising results, please refer to [link to relevant documentation]. Figure 2 , Figure 2 The diagram shows the point cloud data preprocessing result in one embodiment of the present invention. Here, 1 represents a schematic diagram of the tunnel surrounding rock point cloud data, and 2 represents a schematic diagram of the tunnel surrounding rock point cloud noise data. This step enables non-contact, high-density digital acquisition of the tunnel structural surface, avoiding the subjective bias and operational risks of manual measurement, while fully capturing the spatial geometric information of the tunnel structural surface. This provides accurate and comprehensive data support for subsequent steps such as main body extension direction analysis, slicing processing, and feature extraction, ensuring the reliability and engineering practicality of the final structural surface spatial orientation identification result.

[0025] Furthermore, in step S2, due to the scattered distribution of the original 3D point cloud data and the lack of a unified analytical benchmark, direct processing is difficult to accurately focus on the structural features of the tunnel. Moreover, relying on manually specified axes cannot adapt to the actual changes in tunnel morphology. Analyzing the main extension direction provides a precise directional basis for subsequent slicing, ensuring that the slices are consistent with the tunnel's extension trend. Specifically, firstly, the spatial location information of each 3D point cloud data is analyzed to construct a spatial location covariance matrix. Then, principal component analysis is used to perform eigenvalue decomposition on this matrix, extracting the principal eigenvector that represents the maximum dispersion of the point cloud spatial distribution. Finally, this principal eigenvector is used as the tunnel's main extension direction to determine the slicing direction information. The advantage of this step is that it achieves automatic extraction and geometric alignment of the tunnel axis direction without manual intervention. It adapts to the actual changes in tunnel morphology, laying the foundation for subsequent equidistant slicing and 2D projection transformation, making subsequent data processing more targeted and orderly, and improving the accuracy and efficiency of overall structural surface identification.

[0026] Preferably, to improve the accuracy of the original point cloud data, it is necessary to first correct and purify the original point cloud data based on the geometric environment characteristics of the tunnel, and then purify and calibrate the original point cloud data based on prior environmental constraints. The correction and purification process based on the geometric environment characteristics of the tunnel mainly includes: ① converting the scanner's independent coordinate system to the tunnel engineering coordinate system; ② extracting the region of interest (ROI), extracting the point cloud data of the target area by selecting coordinate bounding boxes based on the joint development range of previous geological exploration; ③ setting an appropriate voxel size based on the joint surface texture characteristics, reducing the number of original point clouds to reduce the subsequent computational load while preserving the joint surface undulation characteristics.

[0027] To ensure the accuracy of subsequent plane fitting, this invention utilizes the design axis and theoretical cross-section of the tunnel project as prior environmental constraints to perform directional removal and spatial calibration of "non-rock mass" features from the original point cloud data, removing irrelevant and noisy points. Preferably, to achieve purification and calibration of the original point cloud data based on prior environmental constraints, firstly, the original point cloud data is centered, calculating the arithmetic mean of the three-dimensional coordinates of all discrete points to obtain the geometric centroid of the original point cloud data; this centroid coordinate is then subtracted from all point cloud coordinates to construct a decentralized coordinate dataset with the centroid as the origin, eliminating the impact of excessively large absolute coordinate values ​​on the accuracy of subsequent calculations. Next, a full spatial position covariance matrix is ​​constructed using the decentralized dataset, and eigenvalue decomposition is performed on this real symmetric matrix using a numerical iteration method to extract three eigenvalues ​​and their corresponding unit eigenvectors. The eigenvector corresponding to the largest eigenvalue is selected; this vector represents the direction with the largest spatial dispersion of the point cloud, which is determined to be the direction of the main axis extending the tunnel as a whole, and this main axis direction is also the subsequent slicing direction.

[0028] Furthermore, in step S3, due to the large scale and complex spatial distribution of the 3D point cloud data, directly performing feature extraction and deviation analysis would increase computational difficulty and reduce processing efficiency. Slicing, however, can transform the 3D data into ordered 2D data, providing a clear and regular foundation for subsequent analysis. Specifically, using the axial direction indicated by the slicing direction information as a reference, all 3D point cloud data are sliced ​​at equal intervals greater than the average density of the 3D point cloud data to obtain point cloud slice data. Then, based on the axial direction, a 2D projection transformation is performed on all point cloud slice data to finally obtain 2D tunnel slice data. This process discretizes the large-scale, scattered 3D point cloud into a continuous set of 2D slices, simplifying data processing complexity, reducing computational resource consumption, and making the geometric features of the point cloud easier to represent on a 2D plane. This provides an ordered and unified data structure for subsequent spatial distribution feature extraction and spatial deviation field processing, improving the efficiency and accuracy of the overall processing flow.

[0029] Preferably, in the process of slicing the 3D point cloud data, firstly, based on the extracted principal axis direction, the principal axis direction is converted into a vector representation to obtain the principal axis vector. Secondly, a rotation transformation matrix is ​​constructed from the original scanning coordinate system to the local engineering coordinate system of the tunnel. Subsequently, according to the engineering design accuracy requirements (e.g., 0.5 meters to 1 meter), equally spaced slice intervals are set along the Z-axis direction of the new coordinate system. All transformed original point cloud data are traversed, and their Z-axis coordinate values ​​are mapped to the corresponding slice index, thereby discretizing the large-scale scattered 3D point cloud into a series of continuous 2D tunnel slice data, i.e., a set of 2D spatial slices. Specifically, the rotation transformation matrix is ​​determined by setting the principal axis vector as the Z-axis of the new coordinate system and using the Schmidt orthogonalization method to construct a plane perpendicular to the Z-axis as the XY cross-sectional plane, thus determining the rotation matrix parameters. This matrix is ​​used to perform rigid body rotation transformation on all point clouds to achieve automatic correction of the point cloud attitude, making the tunnel axis perpendicular to the XY plane.

[0030] Furthermore, in step S4, since the two-dimensional tunnel slice data is still presented in a scattered coordinate form, lacking a unified feature representation benchmark, it is difficult to accurately quantify the deviation between actual points and theoretical contours when directly used for subsequent deviation analysis. Feature extraction can transform the scattered data into structured features that can be used for comparative analysis, laying the foundation for subsequent deviation field processing. Specifically, firstly, the spatial coordinate information of each two-dimensional tunnel slice data is extracted, and then this spatial coordinate information is converted into polar coordinate information. That is, with the geometric center of each slice as the pole, the polar angle and polar radius of each point within the slice are calculated. Finally, based on this polar coordinate information, features that reflect the spatial distribution pattern of the point cloud are extracted to obtain spatial distribution features. The advantage of doing so is that a unified polar coordinate description system is established, transforming complex planar geometric morphology analysis into simple polar angle and polar radius analysis, making the spatial distribution pattern of the point cloud easier to capture. At the same time, it provides a standardized data format for subsequent deviation calculation based on the polar radius constraint model, ensuring the accuracy and efficiency of deviation analysis, thereby improving the accuracy of overall structural surface identification.

[0031] Preferably, for each two-dimensional tunnel slice data, the spatial coordinate information is converted into polar coordinate information. The process includes: first, traversing each subset of two-dimensional tunnel slice data. Calculate the arithmetic mean of the coordinates of all points in the subset, and define it as the geometric center of the slice. The pole is dynamically determined based on the slice position, adapting to minor local deflections of the tunnel axis. Next, all points within the 2D tunnel slice data are traversed, and a local polar coordinate system is established with the determined pole as the origin. The Euclidean distance of any point within the slice relative to the pole is calculated as the polar radius, and the angle between the point and the horizontal direction is calculated as the polar angle. This completes the data mapping from the Cartesian coordinate system to the polar coordinate system, providing a unified benchmark for subsequent contour analysis.

[0032] Wherein, any point within the slice polar angle and polar radius The calculation formula is as follows: Furthermore, in step S5, since the spatial distribution characteristics only reflect the geometric state of the actual point cloud and lack a comparison benchmark with the tunnel design outline, it is impossible to distinguish between effective rock wall points and interference points such as construction facilities and deep noise points. However, spatial deviation field processing can quantify the degree of deviation between the actual points and the theoretical outline, laying the foundation for subsequent accurate screening of effective data. Specifically, based on the actual geometric shape of the tunnel, the cross-section is first decomposed into basic geometric primitives such as the arch arc segment and the straight sidewall segment. For each primitive, a polar diameter generation rule is established. Within the full 360-degree range, the minimum polar diameter of each primitive is selected and spliced ​​into a continuous theoretical outline envelope to construct a polar diameter constraint model. Then, using the polar angle of each two-dimensional slice point as an index, the theoretical polar diameter of the corresponding angle in the model is queried. The normal radial deviation is obtained by subtracting the actual polar diameter of the point from the theoretical polar diameter. This calculation is performed on all points to form a spatial deviation result covering the entire field. The advantage of this approach is that it transforms complex three-dimensional spatial positional differences into intuitive single radial distance differences, accurately quantifying the deviation of each point from the design profile. It does not rely on point cloud intensity information or conventional filtering algorithms, but effectively identifies interference points in pure geometric space. This provides a reliable basis for subsequent screening of valid rock face data based on deviation results, significantly improving the accuracy of subsequent plane fitting and dip angle calculations.

[0033] Preferably, the model is based on the extreme radius constraint. The calculated normal radial deviation can be expressed as: Where i represents the geometry of the i-th actual tunnel (e.g., i=1 is the arch, i=2 is the left wall, and i=3 is the right wall). Let this be the distance generating function for the i-th geometric shape in polar coordinates. Let represent the parameter vector of the i-th geometric shape. When the geometric shape contains a curve, (R represents the radius of curvature), when there are straight lines in the geometry, (W represents the width of the opening), and min represents the minimum distance calculated from each geometric shape at the same angle, which constitutes the effective physical boundary of the cross-section.

[0034] Furthermore, in step S6, since the 3D point cloud data contains interference points from internal facilities such as construction trolleys, ventilation pipes, and personnel, as well as external noise points such as noise from deep sensors and invalid reflection points, these interference points will affect the accuracy of subsequent plane fitting. Feature filtering can remove invalid data and retain valid rock wall data, laying the foundation for accurate fitting. Specifically, each 3D point cloud data is first characterized, and then the deviation field distribution histogram is statistically analyzed based on engineering experience. Internal interference thresholds and external noise thresholds are set, and all point cloud data are traversed to perform bidirectional logical judgment. If the normal radial deviation of a point is positive and exceeds the internal interference threshold, it is judged as an internal facility interference point and removed. If the deviation is negative and exceeds the absolute value of the external noise threshold, it is judged as a deep noise point and removed. Finally, the set of points with deviations within the threshold range is retained as the point cloud data to be fitted. The advantage of this approach is that it enables the targeted removal of non-rock mass interference points without relying on point cloud intensity information or conventional filtering algorithms. Data purification is completed in pure geometric space, effectively improving the purity and effectiveness of point cloud data, reducing computational interference in subsequent plane fitting, ensuring the accuracy of the fitting model, and thus providing reliable data support for subsequent dip angle calculations.

[0035] Preferably, in order to perform feature filtering on all 3D point cloud data based on spatial deviation results, the normal radial deviation index of all 3D point cloud data is first defined. Secondly, a discrimination model based on the engineering environment is constructed, and the internal interference threshold is set as follows: The external interference threshold is .

[0036] like This indicates that the measured radius is significantly smaller than the theoretical radius, and the judgment point... Construction facilities (such as trolleys, ventilation ducts, and personnel) that intrude into the tunnel's outline must be removed; if This indicates that the measured point is significantly larger than the theoretical radius, and is therefore determined to be deep sensor noise or an invalid reflection point, and is thus discarded; only those points meeting the requirements are retained. The points are used as valid rock wall data.

[0037] Furthermore, in step S7, since a single fitting method struggles to balance anti-interference and fitting accuracy, using only random sampling methods is prone to parameter bias, and using only the least squares method is susceptible to residual noise. The two-stage fitting method combines the advantages of both methods to accurately obtain the planar model of the tunnel structure, thereby accurately deriving its spatial orientation. Specifically, the first stage employs a random sampling consistency method, randomly selecting three non-collinear points from the point cloud data to be fitted to form a sampling combination. Planar model parameters are calculated, the number of interior points in each model is counted, and reliability is evaluated. After traversing a preset number of sampling iterations, the total number of interior points, average distance, and planarity of the interior point set are considered to select the preliminary optimal model. The second stage performs outlier detection on the interior point set of the preliminary optimal model, removes extreme outliers, and then refits using the least squares method to obtain accurate point cloud fitting results. For details, please refer to [link to relevant documentation]. Figure 3 , Figure 3 The diagram shown is a schematic diagram of the tunnel surrounding rock joint fitting plane in one embodiment of the present invention. Here, 3 refers to the schematic diagram of the tunnel surrounding rock joint fitting plane. The dip angle and dip direction are then calculated using the normal vector of the fitting plane. The dip angle is obtained by converting the angle between the normal vector and the vertical direction. The dip direction is determined by the angle between the normal vector and the true north direction after horizontal projection calibration. Finally, the dip angle and dip direction data are integrated to obtain the spatial orientation of the structural surface. For details, please refer to [link to relevant documentation]. Figure 4 , Figure 4 The diagram shows the spatial orientation of tunnel surrounding rock joints in one embodiment of the present invention, where 4 refers to the normal vector direction and 5 refers to the dip and dip angle directions. The advantage of this approach is that it effectively avoids the shortcomings of a single fitting method. The first-stage fitting eliminates most interference factors, while the second-stage fitting improves model accuracy, making the fitting results more closely match the actual structural surface morphology. The calculated dip and dip angle data are highly accurate and repeatable, providing reliable geological parameter support for tunnel surrounding rock stability assessment and support design.

[0038] Preferably, a random sampling consensus method is used for plane fitting. First, a minimum number of points (at least 3 points are required for plane fitting) are randomly selected from the point cloud data to be fitted to determine the plane model. Based on these three points, the plane model parameters (a, b, c, d) are calculated to satisfy the plane equation: Calculate the distance from each point in the point cloud data to be fitted to the plane, and count the points that meet the distance threshold. The number of interior points, where the distance formula is: in, Let be the coordinates of the i-th point in the point cloud data to be fitted.

[0039] The key parameter for plane fitting is the distance threshold. This is used to determine whether a point belongs to the fitted plane. The final planar model is represented by a coefficient vector. Where n is the normal vector of the plane.

[0040] Preferably, the plane normal vector is normalized: Calculate the tilt angle of the 3D point cloud data based on the normalized plane normal vector: The tilt angle is the angle between the horizontal projection of the plane's normal vector and the true north direction. Another embodiment of the present invention provides a tunnel structure surface identification device; for details, please refer to [link to relevant documentation]. Figure 5 , Figure 5 The diagram shown illustrates the structure of a tunnel structure surface identification device according to one embodiment of the present invention. The device includes: Data acquisition module 11 is used to acquire all three-dimensional point cloud data of the tunnel structure surface; Direction analysis module 12 is used to perform main body extension direction analysis on all the three-dimensional point cloud data to obtain slice direction information; Data slicing module 13 is used to slice all the three-dimensional point cloud data based on the slicing direction information to obtain two-dimensional tunnel slice data; Feature extraction module 14 is used to perform feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features; The deviation processing module 15 is used to perform spatial deviation field processing on the spatial distribution characteristics based on a pre-constructed extreme diameter constraint model to obtain spatial deviation results. The extreme diameter constraint model is constructed from the analysis results of the actual geometry of the tunnel structure surface, and the spatial deviation field processing process is configured to analyze the degree of theoretical and actual deviation of the spatial distribution characteristics of the tunnel structure surface. Data filtering module 16 is used to perform feature filtering on all the three-dimensional point cloud data based on the spatial deviation results to obtain the point cloud data to be fitted. The spatial orientation determination module 17 is used to perform two-stage fitting on all the point cloud data to be fitted, and determine the spatial orientation of the tunnel structure surface based on the two-stage fitting results.

[0041] Another embodiment of the present invention provides a tunnel structure surface identification device; for details, please refer to [link to relevant documentation]. Figure 6 , Figure 6The diagram shown illustrates a structural block diagram of a tunnel structure surface identification device according to one embodiment of the present invention. The tunnel structure surface identification device provided in this embodiment includes a processor 21, a memory 22, and a computer program stored in the memory 22 and configured to be executed by the processor 21. When the processor 21 executes the computer program, it implements the steps described in the above embodiment of the tunnel structure surface identification method, for example... Figure 1 Steps S1 to S7 as described in the document.

[0042] For example, the computer program may be divided into one or more modules, which are stored in the memory 22 and executed by the processor 21 to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the tunnel structure face recognition device.

[0043] The tunnel structure surface identification device may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will understand that the schematic diagram is merely an example of a tunnel structure surface identification device and does not constitute a limitation on such a device. It may include more or fewer components than illustrated, or combine certain components, or use different components. For example, the tunnel structure surface identification device may also include input / output devices, network access devices, buses, etc.

[0044] The processor 21 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 21 is the control center of the tunnel structure face recognition device, connecting all parts of the device via various interfaces and lines.

[0045] The memory 22 can be used to store the computer program and / or modules. The processor 21 implements various functions of the tunnel structure face recognition device by running or executing the computer program and / or modules stored in the memory 22 and calling the data stored in the memory 22. The memory 22 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory 22 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0046] If the integrated module of the tunnel structure surface identification device is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0047] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0048] Accordingly, embodiments of the present invention provide a computer-readable storage medium, the computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform steps in a tunnel structure surface identification method as described in the above embodiments, for example... Figure 1 Steps S1 to S7 as described in the document.

[0049] Compared with existing technologies, the beneficial effects of the embodiments of the present invention are at least one of the following: The present invention achieves non-contact, high-density, and digital acquisition of tunnel surrounding rock by obtaining three-dimensional point cloud data of tunnel structural surfaces, solving the problems of low efficiency and incomplete coverage of traditional manual measurement; The present invention achieves automatic extraction and geometric alignment of tunnel axis direction by analyzing the main extension direction of three-dimensional point cloud data and determining the slicing direction, solving the problem of relying on manual specification or design of the axis and being unable to adapt to changes in actual tunnel morphology; The present invention achieves structured expression of point cloud geometric information by converting three-dimensional point cloud data into two-dimensional slice dataset and extracting features from it, providing a unified comparison benchmark for subsequent deviation analysis; The present invention performs spatial deviation field processing on spatial distribution features based on the polar radius constraint model, realizing a quantitative assessment of the degree of geometric deviation between each point and the design contour; The present invention performs feature screening based on spatial deviation results and performs two-stage fitting on the screened results, realizing automatic identification and removal of non-rock mass interference points, solving the problem that single fitting methods are easily affected by noise, and providing a reliable basis for obtaining accurate spatial orientation of tunnel structural surfaces.

[0050] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for identifying tunnel structural surfaces, characterized in that, include: Acquire all 3D point cloud data of the tunnel structure surface; Perform main body extension direction analysis on all the aforementioned 3D point cloud data to obtain slice direction information; Based on the slice direction information, slice processing is performed on all the three-dimensional point cloud data to obtain two-dimensional tunnel slice data; Feature extraction processing is performed on all the two-dimensional tunnel slice data to obtain spatial distribution features; Based on the pre-constructed extreme diameter constraint model, the spatial distribution characteristics are processed by spatial deviation field to obtain spatial deviation results. The extreme diameter constraint model is constructed from the analysis results of the actual geometry of the tunnel structure surface, and the spatial deviation field processing is configured to analyze the degree of theoretical and actual deviation of the spatial distribution characteristics of the tunnel structure surface. Specifically, based on the actual geometric shape, all basic geometric primitives of the tunnel structure surface are determined, and all the basic geometric primitives are spliced ​​together to obtain the contour envelope. Based on the contour envelope, the polar radius constraint model is constructed. Then, using the polar angle of the two-dimensional slice point in each two-dimensional tunnel slice data as an index, the theoretical polar radius of the corresponding angle in the polar radius constraint model is queried. The actual polar radius of the two-dimensional slice point is subtracted from the theoretical polar radius to obtain the normal radial deviation. This calculation is performed on all the two-dimensional slice points to form the spatial deviation result covering the entire field. Based on the spatial deviation results, feature filtering is performed on all the three-dimensional point cloud data to obtain the point cloud data to be fitted. A two-stage fitting is performed on all the point cloud data to be fitted, and the spatial orientation of the tunnel structure surface is determined based on the two-stage fitting results.

2. The method for identifying tunnel structural surfaces as described in claim 1, characterized in that, The step of performing main body extension direction analysis on all the three-dimensional point cloud data to obtain slice direction information includes: The spatial location information corresponding to each extracted 3D point cloud data is analyzed to obtain the spatial location covariance matrix. Based on the principal component analysis method, the spatial location covariance matrix is ​​subjected to eigenvalue decomposition to obtain the principal eigenvectors; The slice direction information is determined based on the main feature vector.

3. The method for identifying tunnel structural surfaces as described in claim 1, characterized in that, The step of slicing all the three-dimensional point cloud data based on the slicing direction information to obtain two-dimensional tunnel slice data includes: Using the axial direction indicated by the slice direction information as a reference, all the three-dimensional point cloud data are sliced ​​at equal intervals to obtain point cloud slice data, wherein the interval of the equal interval slice processing is greater than the average density of the three-dimensional point cloud data. Based on the axis direction, a two-dimensional projection transformation is performed on all the point cloud slice data to obtain the corresponding two-dimensional tunnel slice data.

4. The method for identifying tunnel structural surfaces as described in claim 1, characterized in that, The step of performing feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features includes: Extract the spatial coordinate information of all the two-dimensional tunnel slice data; The spatial coordinate information is transformed into polar coordinates to obtain the polar coordinate information corresponding to each of the two-dimensional tunnel slice data. Based on the polar coordinate information, feature extraction processing is performed on the corresponding two-dimensional tunnel slice data to obtain the spatial distribution features.

5. The method for identifying tunnel structural surfaces as described in claim 1, characterized in that, The pre-constructed polar radius constraint model performs spatial deviation field processing on the spatial distribution characteristics to obtain spatial deviation results, including: Based on the aforementioned extreme diameter constraint model, spatial characteristic analysis is performed on the tunnel structural surface to obtain theoretical spatial distribution characteristics; The spatial distribution characteristics and the theoretical spatial distribution characteristics are subjected to the spatial deviation field processing to obtain the spatial deviation result.

6. The method for identifying tunnel structural surfaces as described in claim 1, characterized in that, The step of performing feature filtering on all the three-dimensional point cloud data based on the spatial deviation results to obtain the point cloud data to be fitted includes: Each of the three-dimensional point cloud data is subjected to feature labeling processing to obtain feature labeling results; Based on the spatial deviation results, all the feature labeling results are filtered to obtain the point cloud data to be fitted.

7. The method for identifying tunnel structural surfaces as described in claim 1, characterized in that, The step of performing a two-stage fitting on all the point cloud data to be fitted, and determining the spatial orientation of the tunnel structure surface based on the two-stage fitting results, includes: Based on a random sampling algorithm, the point cloud data to be fitted is subjected to a first-stage fitting process to obtain an initial fitting result. Based on the least squares method, the initial fitting result is subjected to a second-stage fitting process to obtain the point cloud fitting result; Spatial orientation analysis is performed on the point cloud fitting results to obtain the dip and tilt angle data of the tunnel structure surface; By integrating the dip data and the tilt angle data, the spatial orientation of the tunnel structure surface is obtained.

8. A tunnel structure surface identification device, characterized in that, include: The data acquisition module is used to acquire all three-dimensional point cloud data of the tunnel structure surface; The orientation analysis module is used to perform main extension orientation analysis on all the three-dimensional point cloud data to obtain slice orientation information; The data slicing module is used to slice all the three-dimensional point cloud data based on the slicing direction information to obtain two-dimensional tunnel slice data; The feature extraction module is used to perform feature extraction processing on all the two-dimensional tunnel slice data to obtain spatial distribution features; The deviation processing module is used to perform spatial deviation field processing on the spatial distribution characteristics based on a pre-constructed extreme diameter constraint model to obtain spatial deviation results. The extreme diameter constraint model is constructed from the analysis results of the actual geometry of the tunnel structure surface, and the spatial deviation field processing process is configured to analyze the degree of theoretical and actual deviation of the spatial distribution characteristics of the tunnel structure surface. Specifically, based on the actual geometric shape, all basic geometric primitives of the tunnel structure surface are determined, and all the basic geometric primitives are spliced ​​together to obtain the contour envelope. Based on the contour envelope, the polar radius constraint model is constructed. Then, using the polar angle of the two-dimensional slice point in each two-dimensional tunnel slice data as an index, the theoretical polar radius of the corresponding angle in the polar radius constraint model is queried. The actual polar radius of the two-dimensional slice point is subtracted from the theoretical polar radius to obtain the normal radial deviation. This calculation is performed on all the two-dimensional slice points to form the spatial deviation result covering the entire field. The data filtering module is used to perform feature filtering on all the three-dimensional point cloud data based on the spatial deviation results to obtain the point cloud data to be fitted. The spatial orientation determination module is used to perform two-stage fitting on all the point cloud data to be fitted, and to determine the spatial orientation of the tunnel structure surface based on the two-stage fitting results.

9. A tunnel structure surface identification device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements a tunnel structure surface identification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein when the device containing the computer-readable storage medium executes the computer program, it implements a tunnel structure surface identification method as described in any one of claims 1 to 7.

Images