Hierarchical calculation and analysis method for non-uniform three-dimensional deformation of mining roadway

Abstract

The invention belongs to the technical field of mining engineering, discloses a mining roadway non-uniform three-dimensional deformation grading calculation and analysis method, and aims to solve the problems that a traditional point cloud comparison method is low in precision and large in error in a roadway high-deformation and non-uniform-deformation environment, and traditional monitoring wiring is complex and cannot achieve global coverage. According to the technical key points, in a GNSS-free underground environment, a portable three-dimensional laser scanning device integrating SLAM synchronous positioning and a map construction technology is adopted to collect two-stage roadway point cloud data, a centimeter-level precision three-dimensional model is constructed through a pose map optimization algorithm, a two-stage deformation hierarchical optimization algorithm is designed, and a two-stage deformation hierarchical optimization algorithm is designed. In the first stage, a deformation core point is selected based on super voxel segmentation, rough calculation is performed in a cylindrical projection space, and a key deformation area is rapidly identified, and in the second stage, a refined displacement amount is calculated for a point whose displacement exceeds a threshold value through geometric constraint and a centroid method, and the data quality is improved in combination with point cloud denoising, so that high-precision continuous monitoring of the full section of the roadway is realized.

Description

Technical Field

[0001] This invention belongs to the field of mining engineering technology, specifically a method for calculating and analyzing the non-uniform three-dimensional deformation of mining roadways. Background Technology

[0002] With the increasing depth and intensity of coal mining in my country, the deformation of surrounding rock in mining roadways under conditions of strong mining activity, high ground stress, and complex geological conditions is becoming increasingly prominent, seriously affecting the safe production of mines. At present, traditional methods for monitoring the deformation of surrounding rock in roadways mainly rely on equipment such as convergence instruments, total stations, and displacement sensors, and adopt point-to-point or two-dimensional monitoring methods. These methods suffer from problems such as complex wiring, low monitoring frequency, and insufficient spatial coverage, making it difficult to achieve full-section, continuous, and high-precision deformation monitoring of roadways. In recent years, three-dimensional laser scanning technology has been gradually applied to mine roadway monitoring due to its non-contact, high-precision, and high-efficiency characteristics.

[0003] In existing technologies, tunnel deformation monitoring methods based on three-dimensional laser scanning, such as C2C and M3C2, mainly adopt point-to-point or point cloud overall comparison methods, which are difficult to effectively deal with complex situations such as large deformation and non-uniform deformation of the surrounding rock in tunnels. They also suffer from geometric mean weakening effect, resulting in insufficient deformation recognition accuracy and difficulty in accurately capturing key deformation areas. At the same time, traditional monitoring equipment has complex wiring and cumbersome operation, and cannot achieve full-section, continuous, and real-time monitoring of the tunnel. The monitoring efficiency and spatial coverage are limited, making it difficult to meet the high-precision, comprehensive, and real-time requirements for tunnel deformation monitoring under high-intensity mining conditions. Summary of the Invention

[0004] The purpose of this invention is to provide a method for calculating and analyzing the non-uniform three-dimensional deformation of mining roadways, so as to solve the problem of insufficient deformation identification accuracy in existing roadway deformation monitoring methods.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for calculating and analyzing the non-uniform three-dimensional deformation of coal mining roadways, addressing the monitoring needs of non-uniform deformation under strong mining conditions in coal mine roadways, and overcoming the geometric mean weakening effect of traditional point cloud comparison methods, comprising the following steps: Sp1: Point cloud data acquisition. A portable 3D laser scanning device integrating SLAM (Simultaneous Localization and Mapping) technology was used to acquire two phases of point cloud data for the tunnel in an underground environment without GNSS signals. The two phases of point cloud data include the first phase point cloud P. t Second phase point cloud P t+1The portable 3D laser scanning device constructs a 3D model with centimeter-level accuracy through a pose graph optimization algorithm. During the acquisition of point cloud data in both phases, the designed cross-sectional height h of the tunnel is recorded simultaneously, and the diameter D of the local cylindrical neighborhood is preset. D is set according to the stability level of the surrounding rock of the tunnel and the point cloud density, and h is the designed cross-sectional height of the tunnel. By using a portable 3D laser scanning device with integrated SLAM technology, high-precision point cloud data of the tunnel in two phases can be quickly acquired in a complex underground environment without GNSS signals, and a 3D model with centimeter-level accuracy can be constructed. This effectively solves the problems of low accuracy and poor efficiency of traditional measurement methods and provides a reliable data foundation for tunnel deformation monitoring.

[0006] Sp2: Point cloud denoising processing, including primary denoising and secondary restoration. Primary denoising is based on the tunnel axis direction. With point cloud normal vector i vertical relationship i =0, calculate the deviation θ of the normal vector. i If θ i >θ th Then it is determined to be a noise point, where θ th The preset angle tolerance threshold is used; secondary recovery is performed on the discarded point set P. noise With the set of points to be optimized P opt Calculate the point p∈P noise To P opt The minimum Euclidean distance d min（p） = min||P opt,i -P noise,j ||2, if d min（p） Less than the preset distance recovery threshold d recover Then p will be reintegrated into the optimized point set, where d recover Positively correlated with the average spacing of point clouds, primary denoising effectively removes outliers that deviate from the main direction of the roadway, while secondary restoration avoids accidental deletion of real point clouds, significantly improving the quality of point cloud data and ensuring the accuracy and reliability of subsequent deformation analysis.

[0007] Sp3: Deformation grading and quantization includes a coarse calculation stage and a fine calculation stage. In the coarse calculation stage, after point cloud denoising, the diameter D of the local cylindrical neighborhood is preset according to the stability level of the surrounding rock of the roadway and the point cloud density, and the radius r of the cylindrical neighborhood is set to D / 2. At the same time, the height H of the designed cross-section of the roadway is used as the height H of the cylindrical neighborhood. The deformation core point p is selected by super-voxel segmentation. i , with p i Centered on the target, a local point cloud p is obtained with the preset diameter D as the local neighborhood range. k Calculate the covariance matrix C, and take the normal vector corresponding to the smallest eigenvalue of C. , construct Let H be a cylindrical projection space with axis r, radius r, and height H. Project the first and second phase point clouds into this cylindrical space respectively, and calculate the average position H of the two projected point clouds. t and H t+1 The average position difference ΔH = ||H t -H t+1 ||, during the precise calculation stage, when ΔH>ΔH th When the deformation core point p exceeds the threshold, t A dynamic safety threshold ΔH is set based on the allowable deformation of the roadway support structure. th Filtering displacement vectors t→t+1 With normal vector The neighborhood set ψ(p) of the angle ϕ < θ t ), where the acute angle threshold θ is used to constrain the direction consistency between the displacement vector and the normal vector, and the refined displacement d(p) is calculated using the centroid method. t )=||p t - ∑ψ(p t )||2, through the coarse calculation stage, the overall deformation trend of the roadway is quickly identified, and the fine calculation stage performs high-precision displacement quantification of key areas, which effectively overcomes the problem of geometric mean weakening, significantly improves the accuracy and reliability of deformation monitoring, and realizes high-precision hierarchical identification and quantitative analysis of roadway surrounding rock deformation.

[0008] Sp4: Deformation law analysis, based on ΔH and d(p) t Quantify the deformation characteristics of the entire cross section of the roadway, identify the deformation differences caused by uneven distribution of support pressure and the geologically weak sections, accurately identify the deformation differences caused by uneven distribution of support pressure, and precisely locate the geologically weak sections, so as to provide a scientific basis for roadway stability evaluation and support optimization.

[0009] Preferably, the 3D laser scanning equipment includes a laser sensor, an IMU inertial measurement unit, and a computing platform, with a maximum range of ≥190m and a scanning rate of ≥240,000 points / second. The equipment adopts SLAM technology and performs nonlinear optimization on a pose graph containing keyframe nodes, roadway geometric feature nodes, and return node nodes to achieve high-precision registration of point clouds. By integrating the laser sensor, IMU inertial measurement unit, and high-performance computing platform, combined with the SLAM pose graph nonlinear optimization algorithm, high-precision registration and 3D modeling of point cloud data in GNSS-free underground environments are achieved, significantly improving scanning efficiency and model accuracy, and meeting the high-density, high-fidelity data acquisition requirements of complex roadway scenarios.

[0010] Preferably, the angle tolerance threshold θth in Sp2 is adaptively set based on the curvature of the tunnel surface, and the distance recovery threshold drecover Positively correlated with the average spacing of the point cloud, the angle tolerance threshold θ is adaptively set based on the curvature of the tunnel surface. th The distance recovery threshold d is dynamically adjusted based on the average spacing of the point cloud. recover This effectively improves the denoising algorithm's adaptability to different geometric feature regions, significantly enhances the authenticity and integrity of point cloud data, and provides a high-quality data foundation for subsequent deformation analysis.

[0011] Preferably, the normal vector The calculation must satisfy: Where k is the number of local point clouds, Using the centroid coordinates, the normal vector is solved by the minimum eigenvalue of the local point cloud covariance matrix, which effectively improves the accuracy and stability of the normal vector estimation, ensuring that local geometric features can still be accurately captured on the surface of complex tunnels, and providing a reliable geometric basis for subsequent deformation analysis.

[0012] Preferably, the height H of the cylindrical projection space is taken as the design cross-sectional height h of the tunnel, and the radius r is taken as half of the preset diameter D. The preset diameter D is set according to the local point cloud density. By setting the height H of the cylindrical projection space to cover the design cross-sectional height range of the tunnel, and matching the radius r with the local point cloud distribution density, it can be effectively ensured that the projection area completely covers the key deformation area of ​​the tunnel, improve the accuracy and representativeness of deformation calculation, and achieve accurate capture of the deformation characteristics of the entire cross-section of the tunnel.

[0013] Preferably, the included angle threshold θ is an acute angle, used to constrain the direction consistency between the displacement vector and the normal vector, and the dynamic safety threshold ΔH th Based on the stability requirements of the surrounding rock in the roadway, by setting the included angle threshold θ to an acute angle, the direction consistency between the displacement vector and the normal vector is effectively constrained, ensuring the spatial accuracy of deformation calculation. Simultaneously, the dynamic safety threshold ΔH... th Based on the allowable deformation of the roadway support structure, it is possible to accurately identify deformation areas that exceed the safety limit, thereby improving the safety and engineering applicability of deformation monitoring.

[0014] Preferably, in Sp4, the deformation difference manifests as the displacement difference in different areas of the roadway cross-section. The geologically weak section is identified by the abrupt change characteristics of the spatial gradient of deformation. By analyzing the deformation difference characteristics, the law that the displacement of the small coal pillar side is significantly greater than that of the solid coal side and the displacement of the floor is significantly greater than that of the two sides is clearly identified. Combined with the abrupt change characteristics of the deformation gradient, the geologically weak section is accurately located, effectively revealing the uneven distribution of support pressure and the difference in surrounding rock stability, providing a reliable basis for roadway support optimization and risk prevention and control.

[0015] Preferably, the search range of the neighborhood point set ψ(pt) is adapted to the point cloud resolution to ensure the completeness of the displacement vector selection. t The search range is adapted to the point cloud resolution to ensure the integrity and accuracy of displacement vector selection, effectively avoid missing key deformation information, and improve the reliability and accuracy of deformation quantization in the fine calculation stage.

[0016] Preferably, the deformation output by the coarse calculation stage and the fine calculation stage meets the requirement that both the average absolute error and the relative error are lower than the preset accuracy threshold, which significantly improves the deformation monitoring accuracy. The average absolute error is controlled within 0.0483 meters and the relative error does not exceed 11.28%, effectively meeting the engineering application's need for high-precision deformation quantification.

[0017] The beneficial effects of this invention are as follows: This invention proposes a graded optimization analysis method for deformation in mining roadways based on 3D laser scanning. By utilizing a portable scanning device integrating SLAM technology, high-precision 3D modeling of the roadway is achieved even in GNSS-free environments, solving the problems of complex wiring, low efficiency, and incomplete coverage associated with traditional monitoring methods. A two-stage point cloud denoising strategy is employed to effectively remove noise and restore the true point cloud, improving data quality. An innovative Two-Stage deformation grading and quantification algorithm is constructed, combining coarse calculation for rapid identification of deformation areas with fine calculation for high-precision quantification of displacement. This overcomes the accuracy challenges posed by geometric mean weakening and non-uniform deformation, significantly improving the accuracy and reliability of deformation monitoring. By analyzing deformation difference characteristics, uneven distribution of support pressure and geologically weak sections are accurately identified, revealing the spatiotemporal evolution law of roadway surrounding rock deformation. Compared with traditional methods, this invention exhibits significant advantages in accuracy, adaptability, and stability, with an average absolute error controlled within 0.0483 meters and a relative error not exceeding 11.28%. It provides reliable data support and scientific basis for roadway stability evaluation, support optimization, and safety early warning, and has broad engineering application prospects. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the hierarchical optimization analysis method of the present invention; Figure 2 This is a schematic diagram of the pose graph SLAM structure of the present invention; Figure 3 This is a schematic diagram illustrating the calculation of the tunnel point cloud normal vector according to the present invention; Figure 4 This is a schematic diagram of the rough calculation of tunnel deformation according to the present invention; Figure 5 This is a schematic diagram of the core point of tunnel deformation in this invention; Figure 6 This is a schematic diagram illustrating the refined calculation of point cloud tunnel deformation according to the present invention. Detailed Implementation

[0019] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Example: like Figures 1 to 6 As shown, this invention provides a method for calculating and analyzing the non-uniform three-dimensional deformation of mining roadways. Addressing the monitoring needs of non-uniform deformation under strong mining conditions in coal mine roadways, it overcomes the geometric mean weakening effect of traditional point cloud comparison methods. The method includes the following steps: Point cloud data acquisition was conducted using a portable 3D laser scanning device integrating SLAM (Simultaneous Localization and Mapping) technology. Two phases of point cloud data were collected in an underground environment without GNSS signals. The two phases of point cloud data include the first phase point cloud P. t Second phase point cloud P t+1 The portable 3D laser scanning device constructs a 3D model with centimeter-level accuracy through a pose graph optimization algorithm. During the acquisition of point cloud data in both phases, the designed cross-sectional height h of the tunnel is recorded simultaneously, and the diameter D of the local cylindrical neighborhood is preset. D is set according to the stability level of the surrounding rock of the tunnel and the point cloud density, and h is the designed cross-sectional height of the tunnel. By using a portable 3D laser scanning device with integrated SLAM technology, high-precision point cloud data of the tunnel in two phases can be quickly acquired in a complex underground environment without GNSS signals, and a 3D model with centimeter-level accuracy can be constructed. This effectively solves the problems of low accuracy and poor efficiency of traditional measurement methods and provides a reliable data foundation for tunnel deformation monitoring.

[0021] Point cloud denoising processing includes primary denoising and secondary restoration. Primary denoising is based on the tunnel axis direction. With point cloud normal vector i vertical relationship i =0, calculate the deviation θ of the normal vector. i If θ i >θ th Then it is determined to be a noise point, where θ th The preset angle tolerance threshold is used; secondary recovery is performed on the discarded point set P. noise With the set of points to be optimized P opt Calculate the point p∈P noise To P opt The minimum Euclidean distance d min（p） = min||Popt,i -P noise,j ||2, if d min（p） Less than the preset distance recovery threshold d recover Then p will be reintegrated into the optimized point set, where d recover Positively correlated with the average spacing of point clouds, primary denoising effectively removes outliers that deviate from the main direction of the roadway, while secondary restoration avoids accidental deletion of real point clouds, significantly improving the quality of point cloud data and ensuring the accuracy and reliability of subsequent deformation analysis.

[0022] Deformation grading and quantification includes a coarse calculation stage and a fine calculation stage. In the coarse calculation stage, after point cloud denoising, the diameter D of the local cylindrical neighborhood is preset according to the stability level of the surrounding rock of the roadway and the point cloud density, and the radius r of the cylindrical neighborhood is set to D / 2. At the same time, the height H of the designed cross-section of the roadway is used as the height H of the cylindrical neighborhood. The deformation core point p is selected by super-voxel segmentation. i , with p i Centered on the target, a local point cloud p is obtained with the preset diameter D as the local neighborhood range. k Calculate the covariance matrix C, and take the normal vector corresponding to the smallest eigenvalue of C. , construct Let H be a cylindrical projection space with axis r, radius r, and height H. Project the first and second phase point clouds into this cylindrical space respectively, and calculate the average position H of the two projected point clouds. t and H t+1 The average position difference ΔH = ||H t -H t+1 ||, during the precise calculation stage, when ΔH>ΔH th When the deformation core point p exceeds the threshold, t A dynamic safety threshold ΔH is set based on the allowable deformation of the roadway support structure. th Filtering displacement vectors t→t+1 With normal vector The neighborhood set ψ(p) of the angle ϕ < θ t ), where the acute angle threshold θ is used to constrain the direction consistency between the displacement vector and the normal vector, and the refined displacement d(p) is calculated using the centroid method. t )=||p t - ∑ψ(p t )||2, through the coarse calculation stage, the overall deformation trend of the roadway is quickly identified, and the fine calculation stage performs high-precision displacement quantification of key areas, which effectively overcomes the problem of geometric mean weakening, significantly improves the accuracy and reliability of deformation monitoring, and realizes high-precision hierarchical identification and quantitative analysis of roadway surrounding rock deformation.

[0023] Deformation law analysis, based on ΔH and d(p)t Quantify the deformation characteristics of the entire cross section of the roadway, identify the deformation differences caused by uneven distribution of support pressure and the geologically weak sections, accurately identify the deformation differences caused by uneven distribution of support pressure, and precisely locate the geologically weak sections, so as to provide a scientific basis for roadway stability evaluation and support optimization.

[0024] Specifically, the 3D laser scanning equipment includes a laser sensor, an IMU (Inertial Measurement Unit), and a computing platform. It has a maximum range of ≥190m and a scanning rate of ≥240,000 points / second. The equipment employs SLAM technology, using a pose graph containing keyframe nodes, roadway geometric feature nodes, and return node points for nonlinear optimization to achieve high-precision registration of point clouds. By integrating the laser sensor, IMU, and high-performance computing platform, combined with the SLAM pose graph nonlinear optimization algorithm, it achieves high-precision registration and 3D modeling of point cloud data in GNSS-free underground environments, significantly improving scanning efficiency and model accuracy, and meeting the high-density, high-fidelity data acquisition requirements of complex roadway scenarios.

[0025] Specifically, in Sp2, the angle tolerance threshold θth is adaptively set based on the curvature of the tunnel surface, and the distance recovery threshold d... recover Positively correlated with the average spacing of the point cloud, the angle tolerance threshold θ is adaptively set based on the curvature of the tunnel surface. th The distance recovery threshold d is dynamically adjusted based on the average spacing of the point cloud. recover This effectively improves the denoising algorithm's adaptability to different geometric feature regions, significantly enhances the authenticity and integrity of point cloud data, and provides a high-quality data foundation for subsequent deformation analysis.

[0026] Specifically, the normal vector The calculation must satisfy: Where k is the number of local point clouds, Using the centroid coordinates, the normal vector is solved by the minimum eigenvalue of the local point cloud covariance matrix, which effectively improves the accuracy and stability of the normal vector estimation, ensuring that local geometric features can still be accurately captured on the surface of complex tunnels, and providing a reliable geometric basis for subsequent deformation analysis.

[0027] Specifically, the height H of the cylindrical projection space is taken as the design cross-sectional height h of the tunnel, and the radius r is taken as half of the preset diameter D. The preset diameter D is set according to the local point cloud density. By setting the height H of the cylindrical projection space to cover the design cross-sectional height range of the tunnel, and matching the radius r with the local point cloud distribution density, it can be effectively ensured that the projection area completely covers the key deformation area of ​​the tunnel, improve the accuracy and representativeness of deformation calculation, and achieve accurate capture of the deformation characteristics of the entire cross-section of the tunnel.

[0028] Specifically, the included angle threshold θ is an acute angle used to constrain the direction consistency between the displacement vector and the normal vector, and the dynamic safety threshold ΔH th Based on the stability requirements of the surrounding rock in the roadway, by setting the included angle threshold θ to an acute angle, the direction consistency between the displacement vector and the normal vector is effectively constrained, ensuring the spatial accuracy of deformation calculation. Simultaneously, the dynamic safety threshold ΔH... th Based on the allowable deformation of the roadway support structure, it is possible to accurately identify deformation areas that exceed the safety limit, thereby improving the safety and engineering applicability of deformation monitoring.

[0029] Specifically, in Sp4, the deformation difference manifests as the displacement difference in different areas of the roadway cross section. The geologically weak section is identified by the abrupt change characteristics of the spatial gradient of deformation. By analyzing the deformation difference characteristics, the law that the displacement of the small coal pillar side is significantly greater than that of the solid coal side and the displacement of the floor is significantly greater than that of the two sides is clearly identified. Combined with the abrupt change characteristics of the deformation gradient, the geologically weak section is accurately located, effectively revealing the uneven distribution of support pressure and the difference in the stability of the surrounding rock, providing a reliable basis for roadway support optimization and risk prevention and control.

[0030] Specifically, the search range of the neighborhood point set ψ(pt) is adapted to the point cloud resolution to ensure the completeness of the displacement vector selection. t The search range is adapted to the point cloud resolution to ensure the integrity and accuracy of displacement vector selection, effectively avoid missing key deformation information, and improve the reliability and accuracy of deformation quantization in the fine calculation stage.

[0031] Specifically, the deformation output by the coarse calculation stage and the fine calculation stage meets the requirement that both the average absolute error and the relative error are lower than the preset accuracy threshold, which significantly improves the deformation monitoring accuracy. The average absolute error is controlled within 0.0483 meters and the relative error does not exceed 11.28%, effectively meeting the engineering application's need for high-precision deformation quantification.

[0032] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0033] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for calculating and analyzing the non-uniform three-dimensional deformation of coal mine roadways, addressing the monitoring needs of non-uniform deformation under strong mining conditions in coal mine roadways, and overcoming the geometric mean weakening effect of traditional point cloud comparison methods, characterized in that... Includes the following steps: Sp1: Point cloud data acquisition. A portable 3D laser scanning device integrating SLAM (Simultaneous Localization and Mapping) technology was used to acquire two phases of point cloud data for the tunnel in an underground environment without GNSS signals. The two phases of point cloud data include the first phase point cloud P. t Second phase point cloud P t+1 The portable 3D laser scanning device constructs a 3D model with centimeter-level precision through a pose graph optimization algorithm. When collecting point cloud data in the two phases, the design cross-sectional height h of the tunnel is recorded simultaneously, and the diameter D of the local cylindrical neighborhood is preset. D is set according to the stability level of the surrounding rock of the tunnel and the point cloud density, and h is the design cross-sectional height of the tunnel. Sp2: Point cloud denoising processing, including primary denoising and secondary restoration. Primary denoising is based on the tunnel axis direction. With point cloud normal vector i vertical relationship i =0, calculate the deviation θ of the normal vector. i If θ i >θ th Then it is determined to be a noise point, where θ th The preset angle tolerance threshold is used; secondary recovery is performed on the discarded point set P. noise With the set of points to be optimized P opt Calculate the point p∈P noise To P opt The minimum Euclidean distance d min（p） = min||P opt,i -P noise,j ||2, if d min（p） Less than the preset distance recovery threshold d recover Then p will be reintegrated into the optimized point set, where d recover Positively correlated with the average spacing of point clouds; Sp3: Deformation-level quantization, including coarse calculation stage and fine calculation stage: Sp3-1: In the coarse calculation stage, after point cloud denoising, the diameter D of the local cylindrical neighborhood is preset based on the stability level of the surrounding rock of the tunnel and the point cloud density, and the radius r of the cylindrical neighborhood is set to D / 2. At the same time, the height H of the designed cross-section of the tunnel is used as the height H of the cylindrical neighborhood, and the deformation core point p is selected by super-voxel segmentation. i , with p i Centered on the target, a local point cloud p is obtained with the preset diameter D as the local neighborhood range. k Calculate the covariance matrix C, and take the normal vector corresponding to the smallest eigenvalue of C. , construct Let H be a cylindrical projection space with axis r, radius r, and height H. Project the first and second phase point clouds into this cylindrical space respectively, and calculate the average position H of the two projected point clouds. t and H t+1 The average position difference ΔH = ||H t -H t+1 ||; Sp3-2: Fine calculation stage, when ΔH>ΔH th When the deformation core point p exceeds the threshold, t A dynamic safety threshold ΔH is set based on the allowable deformation of the roadway support structure. th Filtering displacement vectors t→t+1 With normal vector The neighborhood set ψ(p) of the angle ϕ < θ t ), where the acute angle threshold θ is used to constrain the direction consistency between the displacement vector and the normal vector, and the refined displacement d(p) is calculated using the centroid method. t )=||p t - ∑ψ(p t )||2; Sp4: Deformation law analysis, based on ΔH and d(p) t Quantify the deformation characteristics of the entire cross-section of the tunnel to identify deformation differences caused by uneven distribution of support pressure and geologically weak sections.

2. The hierarchical calculation and analysis method according to claim 1, characterized in that: The three-dimensional laser scanning equipment includes a laser sensor, an IMU inertial measurement unit, and a computing platform. The maximum range is ≥190m, and the scanning rate is ≥240,000 points / second. The equipment adopts SLAM technology and performs nonlinear optimization through a pose graph containing key frame nodes, roadway geometric feature nodes, and return node nodes to achieve high-precision registration of point clouds.

3. The hierarchical calculation and analysis method according to claim 1, characterized in that: In Sp2, the angle tolerance threshold θth is adaptively set based on the curvature of the tunnel surface, and the distance recovery threshold d is also set. recover It is positively correlated with the average spacing of the point cloud.

4. The hierarchical calculation and analysis method according to claim 1, characterized in that: The normal vector The calculation must satisfy: Where k is the number of local point clouds, The coordinates are the centroid coordinates.

5. The hierarchical calculation and analysis method according to claim 1, characterized in that: The height H of the cylindrical projection space is taken as the design cross-sectional height h of the tunnel, and the radius r is taken as half of the preset diameter D, which is set according to the local point cloud density.

6. The hierarchical calculation and analysis method according to claim 1, characterized in that: The included angle threshold θ is an acute angle used to constrain the direction consistency between the displacement vector and the normal vector. The dynamic safety threshold ΔH th The settings are based on the stability requirements of the surrounding rock in the tunnel.

7. The hierarchical calculation and analysis method according to claim 1, characterized in that: In Sp4, the deformation difference is manifested as the displacement difference in different areas of the tunnel cross section, and the geologically weak section is identified by the abrupt change characteristics of the spatial gradient of the deformation.

8. The hierarchical calculation and analysis method according to claim 1, characterized in that: The search range of the neighborhood point set ψ(pt) is adapted to the point cloud resolution to ensure the completeness of the displacement vector selection.

9. The hierarchical calculation and analysis method according to claim 1, characterized in that: The deformation output by the coarse calculation stage and the fine calculation stage together satisfies the condition that both the average absolute error and the relative error are lower than the preset accuracy threshold.

Images