A method for describing rock mass deformation and deformation trend based on 3D laser scanning point cloud data

Abstract

This invention provides a method for describing the deformation and trend of rock mass based on three-dimensional laser scanning point cloud data. The specific steps are as follows: Three-dimensional laser scanning is performed on the rock mass to obtain high-precision point cloud data of the rock mass to be evaluated; the distance between points in the point cloud is calculated, the fitting plane range is defined, and the positive and negative distances between each point and its best-fit plane are calculated; a scatter plot of the relative deformation of each point is drawn. This method for describing the deformation and trend of rock mass utilizes three-dimensional laser scanning to directly process the surface information of the rock mass. It can intuitively reflect the danger zone in a single-period point cloud data, and can also compare the relative deformation data of two periods in the danger zone to determine the deformation development stage and the expansion of the deformation zone. This reduces the process of multi-period point cloud registration, lowers the workload, and makes it more suitable for conveniently and intuitively determining the degree and trend of rock mass deformation and the identification of danger zones.

Description

Technical Field

[0001] This invention relates to the field of slope deformation monitoring technology, and in particular to a method for describing the amount and trend of rock mass deformation based on three-dimensional laser scanning point cloud data. Background Technology

[0002] The deformation development trend and rate of slope, as well as the amount of deformation, are important indicators for evaluating the stability and future development trend of slopes. Slope rock mass deformation monitoring is one of the important means of obtaining slope deformation information. Therefore, how to accurately obtain slope rock mass deformation information is of great significance for evaluating slope stability and disaster prediction.

[0003] The use of 3D laser scanning technology to acquire deformation information has been explored in various fields. Common methods for obtaining deformation using point cloud coordinates include: aligning two point cloud datasets based on points, and then calculating the deformation after aligning the two datasets; or pre-establishing a target model, acquiring point cloud data, and then comparing the point cloud data with the model to determine the deformation. However, using point cloud data for rock mass deformation analysis has the following shortcomings: ① Point cloud alignment for deformation analysis is labor-intensive, inefficient, and produces inconsistent alignment results; the process of comparing multiple datasets is cumbersome. ② Irregular rock mass surfaces make pre-modeling difficult, hindering comparison between data and the model. ③ Single-period point cloud data is insufficient to effectively represent slope rock mass deformation information. ④ Without controllable deformation benchmarks, it is difficult to describe the deformation trend. Summary of the Invention

[0004] The purpose of this invention is to provide a method for describing the deformation amount and trend of rock mass based on three-dimensional laser scanning point cloud data. This method can intuitively identify deformation hazard zones. At the same time, for comparison of data from multiple periods, based on the premise of relative deformation surfaces, it reduces workload, improves efficiency, and reduces alignment difficulty. Furthermore, it adopts image alignment instead of the traditional point cloud alignment method. After image alignment, based on the change in the relative deformation range of the hazard zone, it can intuitively determine the range change of the deformation zone and the development of the deformation trend.

[0005] To achieve the above-mentioned technical features, the objective of this invention is as follows: a method for describing the deformation amount and deformation trend of rock mass based on three-dimensional laser scanning point cloud data, comprising the following steps:

[0006] Step 1: Perform a three-dimensional laser scan of the bank slope rock mass to obtain the spatial coordinate information of the point cloud of the bank slope rock mass to be processed at the observation point;

[0007] Step 2: Use ATLAscan and Geomagic to preprocess the acquired raw point cloud to obtain a first-level point cloud;

[0008] Step 3: Use the first-level point cloud coordinates to determine the best fitting plane; each point has a fitting plane.

[0009] Step 4: Calculate the distance from any point to the best-fit plane;

[0010] Step 5: Distinguish between positive and negative distances obtained in Step 4;

[0011] Step 6: Use RGB for coloring to obtain a point-to-surface distance color map;

[0012] Step 7: Analyze single-period data information;

[0013] Step 8: Obtain the second phase of point cloud data and repeat steps 1 to 7.

[0014] Step 9: Compare the point-to-surface distance color maps of the two periods, and use the highlighted areas to identify easily deformable areas, deformation boundary ranges, and deformation trends.

[0015] Step 10: Compare and analyze the distance distribution characteristic parameters in the danger zone, and obtain the relative distance change based on the changes in data from multiple periods to obtain the relative deformation data.

[0016] The preprocessing of the original point cloud in step 2 includes deleting point clouds outside the study area, point cloud denoising, and resampling.

[0017] The specific operation process in step 3 is as follows:

[0018] Step 3.1, calculate the distance p between any point and other points. ij :

[0019] Calculate point P using the coordinates of each point in the first-level point cloud data. i (i = 1...n) and other points P j The L2 norm of coordinates (j = 1...n, j ≠ i) is calculated as p. ij The result is placed in set D. i (i = 1...n), where n is the total number of points in the first-level point cloud;

[0020] Step 3.2, Selection of points for the fitting plane:

[0021] The range of values ​​for the relative fitting plane is set as d. The value of d depends on the point cloud density and the size of the object under study. If the value of d is too small, there will be too few points for plane fitting, and the deformation of each point will be less sensitive to deformation detection when the deformation is greater than d. If the value of d is larger, the calculation will be slower and the similarity of the fitted planes from different points will be greater, resulting in insensitivity to small distances. For point i(x i ,y i ,z i ), for its corresponding set D i Each element p ijCompared with the d value, if p ij If ≤d, then point j(x) j ,y j ,z j The coordinates of ) are placed in set K. i (i = 1...n);

[0022] Step 3.3: Determine the best-fit plane using the SVD method:

[0023] Calculate set K i The mean values ​​X of the midpoint cloud coordinates X, Y, Z mean ,Y mean Z mean Using set K i Construct a coefficient matrix using all elements:

[0024]

[0025] Perform singular value decomposition on matrix A, A = UDV T Where: D is a diagonal matrix, and U and V are both unitary matrices; the eigenvector corresponding to the minimum singular value is the coefficient vector of the fitting plane, and the fitting plane a i x+b i y+c i z+d i =0, coefficient a i =V(1,3); b i =V(2,3);ci=V(3,3);d i =-[a, b, c]·[X mean Y mean Z mean ].

[0026] The specific calculation process in step 4 is as follows:

[0027] Calculate the distance L(i) from point i to its best-fit plane α:

[0028]

[0029] The specific calculation process in step 5 is as follows:

[0030] The initial point (0, 0, 0) and point i(x) of the three-dimensional laser scan are set. i y i , z i ) coordinates, substitute them into the corresponding equation of point i (a i x+b i y+c i z+d i The result of )*(d) is negative if the distance is negative, and positive otherwise, and is corrected.

[0031] The specific coloring process in step 6 is as follows:

[0032] The calculated distance L is assigned to each point in the point cloud, and RGB coloring is used to obtain a point-to-surface distance color map. Points with closer calculated distances are distinguished by one color, and the area is indicated to be relatively flat, thus reducing the focus. Positive and negative distance values ​​are represented by different colors.

[0033] The specific operation process in step 7 is as follows:

[0034] Based on the point-to-area distance color map, the distance distribution characteristic parameter f is statistically analyzed in the bright color regions. max f mean .

[0035] The present invention has the following beneficial effects:

[0036] 1. This invention can intuitively identify the dangerous deformation zone. At the same time, for comparison of multiple data periods, based on the premise of relative deformation surface, it reduces workload and improves efficiency, reduces alignment difficulty, and can change point cloud alignment to image alignment. After image alignment, based on the change of the relative deformation range of the dangerous zone, the range change of the deformation zone and the development of the deformation trend can be intuitively judged.

[0037] 2. This method makes the following improvements to the current methods for identifying rock mass deformation:

[0038] 1) Previous point cloud deformation description methods often involved aligning the point clouds before analysis or pre-building a model and comparing the point cloud data with the model to obtain deformation. For rock masses, pre-setting a model is labor-intensive and difficult, and the large and irregular distribution of point clouds makes pre-building a model and aligning the point clouds challenging. This paper proposes a relative deformation method, pre-setting a relative deformation surface, which is then used as the reference for subsequent deformation analysis. Furthermore, considering the irregularity of the rock mass, a planar fitting method using the point cloud is proposed, with the fitted plane serving as the relative deformation reference surface.

[0039] 2) Based on the setting of the relative deformation surface, the data acquired in a single period can also be used for deformation analysis to obtain some information about the scanned object.

[0040] 3) Based on the invariance of the overall range during deformation development, invariant points are analyzed as relatively invariant planes. Comparing multiple data periods using the same relative deformation surface makes it easier and more effective to obtain deformation development trends and changes in deformation amount. By comparing the changes in colored deformation areas across multiple data periods to analyze trends, and then comparing the relative deformation amount in areas with significant changes, the amount of change can be determined. This reduces workload. Attached Figure Description

[0041] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0042] Figure 1 This is a schematic diagram of the rock mass deformation description method based on three-dimensional laser scanning point cloud data according to the present invention.

[0043] Figure 2 This diagram illustrates the specific process of calculating the relative distance of a point cloud in a single operation.

[0044] Figure 3 In one embodiment of the present invention, the relative distance of point cloud of the first concrete model is calculated and mapped. The raised areas and the recessed areas can be identified in the map.

[0045] Figure 4 In one embodiment of the present invention, the relative distance of point cloud is calculated and plotted for the second concrete model, and the first deformation is performed on the basis of the first model.

[0046] Figure 5 In one embodiment of the present invention, the relative distance of point cloud is calculated and plotted for the third concrete model, and a second deformation is performed on the basis of the second model. Detailed Implementation

[0047] The embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0048] Example 1:

[0049] To achieve the above-mentioned technical features, the objective of this invention is as follows: a method for describing the deformation amount and deformation trend of rock mass based on three-dimensional laser scanning point cloud data, comprising the following steps:

[0050] Step 1: Perform a three-dimensional laser scan of the bank slope rock mass to obtain the spatial coordinate information of the point cloud of the bank slope rock mass to be processed at the observation point;

[0051] Step 2: Preprocess the acquired raw point cloud using ATLAscan and Geomagic to obtain the first-level point cloud:

[0052] The preprocessing of the original point cloud includes deleting point clouds outside the study area, point cloud denoising, and resampling.

[0053] The ATLAscan is a multi-mode, multi-functional measurement 3D scanner; Geomagic is a 3D design software that combines real-time 3D scanning, 3D point cloud and triangular mesh editing functions with comprehensive CAD modeling, assembly modeling, and 2D drawing output functions.

[0054] Step 3: Determine the best-fit plane using the first-level point cloud coordinates. Each point has a corresponding fitting plane.

[0055] Step 3.1, calculate the distance p between any point and other points.ij :

[0056] Calculate point P using the coordinates of each point in the first-level point cloud data. i (i = 1...n) and other points P j The L2 norm of the coordinates (j = 1...n, j ≠ i) is calculated as p. ij The result is placed in set D. i (i = 1...n), where n is the total number of points in the first-level point cloud;

[0057] Step 3.2, Selection of points for the fitting plane:

[0058] The range of values ​​for the relative fitting plane is set as d. The value of d depends on the point cloud density and the size of the object under study. If the value of d is too small, there will be too few points for plane fitting, and the deformation of each point will be less sensitive to deformation detection when the deformation is greater than d. If the value of d is larger, the calculation will be slower and the similarity of the fitted planes from different points will be greater, resulting in insensitivity to small distances. For point i(x i y i , z i ), for its corresponding set D i Each element p ij Compared with the d value, if p ij If ≤d, then point j(x) j y j , z j The coordinates of ) are placed in set K. i (i = 1...n);

[0059] Step 3.3: Determine the best-fit plane using the SVD method:

[0060] Calculate set K i The mean values ​​X of the midpoint cloud coordinates X, Y, and Z. mean Y mean Z mean Using set K i Construct a coefficient matrix using all elements:

[0061]

[0062] Perform singular value decomposition on matrix A, A = UDV T Where: D is a diagonal matrix, and U and V are both unitary matrices; the eigenvector corresponding to the minimum singular value is the coefficient vector of the fitting plane, and the fitting plane a i x+b i y+c i z+d i =0, coefficient a i =V(1,3); b i =V(2,3);c i =V(3,3);di =-[a, b, c]·[X mean Y mean Z mean ].

[0063] Singular Value Decomposition (SVD) is a cornerstone of machine learning algorithms. In some respects, it is similar to the diagonalization of symmetric or Hermite matrices based on eigenvectors. However, despite their correlation, these two matrix decompositions are significantly different. Spectral analysis is based on the decomposition of eigenvectors of symmetric matrices, while Singular Value Decomposition is a generalization of spectral analysis theory to arbitrary matrices. The importance of SVD in matrix theory is self-evident; it has crucial applications in optimization problems, eigenvalue problems, least power problems, generalized inverse matrices, statistics, image processing, and natural language processing. Therefore, this embodiment employs the SVD method.

[0064] Step 4: Calculate the distance from any point to the best-fit plane:

[0065] Calculate the distance L(i) from point i to its best-fit plane α:

[0066]

[0067] Step 5: Distinguish between positive and negative distances obtained in Step 4:

[0068] The initial point (0,0,0) and point i(x) of the three-dimensional laser scan are set. i ,y i ,z i ) coordinates, substitute them into the corresponding equation of point i (a i x+b i y+c i z+d i The result of )*(d) is negative if the distance is negative, and positive otherwise, and is corrected.

[0069] Step 6: Use RGB for coloring to obtain the point-to-surface distance color map:

[0070] The calculated distance L is assigned to each point in the point cloud, and RGB coloring is used to obtain a point-to-surface distance color map. Points with closer calculated distances are distinguished by one color, and the area is indicated to be relatively flat, thus reducing the focus. Positive and negative distance values ​​are represented by different colors.

[0071] In this embodiment, the calculated distance L is assigned to each point in the point cloud, and RGB coloring is used based on the calculated distance to obtain a point-to-surface distance color map (see attached). Figure 3 Appendix Figure 4In the graph, points that are closer in the calculated distance are represented by a darker color, indicating that the area is relatively flat and can be ignored. Positive and negative distance values ​​are represented by different colors.

[0072] The RGB color mode is an industry color standard that uses variations in the red (R), green (G), and blue (B) color channels and their superposition to obtain a variety of colors. RGB represents the colors of the red, green, and blue channels.

[0073] Step 7, analyze single-period data information:

[0074] Based on the point-to-area distance color map, the distance distribution characteristic parameter f is statistically analyzed in the bright color regions. max f mean wait.

[0075] In this embodiment, the appendix Figure 3 The areas in the air and the protruding sharp corners are clearly marked in red, and the edges of the overall airy areas are clearly circled. The deep recesses are marked in blue.

[0076] Step 8: Obtain the second phase of point cloud data and repeat steps 1 to 7.

[0077] Step 9: Compare the point-to-surface distance color maps of the two periods, and use the highlighted areas to identify easily deformable areas, deformation boundary ranges, and deformation trends.

[0078] In this embodiment, the appendix Figure 4 With appendix Figure 5 In the middle section, there was obvious deformation, with attached... Figure 4 The original area was a flat region; positive values ​​appear at the edges of the deformation, indicating displacement, while negative distance points appear on the other side. (Attached) Figure 5 At the edge of the deformation, it can be observed that the deformation has further expanded. At the same time, a negative value area is more clearly marked at the depression on the other side of the edge, indicating that the observed object at this location has deformed in a direction away from the observation point.

[0079] Step 10: Compare and analyze the distance distribution characteristic parameters in the danger zone, and obtain the relative distance change based on the changes in data from multiple periods to obtain the relative deformation data.

[0080] In this embodiment, according to the appendix Figure 3 With appendix Figure 1 The data comparison showed that the final deformation was about 8cm at the edge of the deformation.

Claims

1. A method for describing the deformation amount and trend of rock mass based on three-dimensional laser scanning point cloud data, characterized in that, Includes the following steps: Step 1: Perform a three-dimensional laser scan of the bank slope rock mass to obtain the spatial coordinate information of the point cloud of the bank slope rock mass to be processed at the observation point; Step 2: Use ATLAscan and Geomagic to preprocess the acquired raw point cloud to obtain a first-level point cloud; Step 3: Use the first-level point cloud coordinates to determine the best fitting plane; each point has a fitting plane. Step 4: Calculate the distance from any point to the best-fit plane; Step 5: Distinguish between positive and negative distances obtained in Step 4; Step 6: Use RGB for coloring to obtain a point-to-surface distance color map; Step 7: Analyze single-period data information; Step 8: Obtain the second phase of point cloud data and repeat steps 1 to 7. Step 9: Compare the point-to-surface distance color maps of the two periods, and use the highlighted areas to identify easily deformable areas, deformation boundary ranges, and deformation trends. Step 10: Compare and analyze the distance distribution characteristic parameters in the danger zone, and obtain the relative distance change based on the changes in data from multiple periods to obtain the relative deformation data; The specific operation process in step 3 is as follows: Step 3.1: Calculate the distance between any point and all other points. p ij : Calculate the points using the coordinates of each point in the first-level point cloud data. P i Compared with other points P j The L2 norm of the coordinate values ​​is calculated as follows: p ij The results are placed in a set D i , n This represents the total number of points in the first-level point cloud. i =1…… n , j =1…… n , j ≠ i ; Step 3.2, Selection of points for the fitting plane: Set the range of values ​​for the relative fitting plane d , d The value depends on the point cloud density and the size of the object under study. d A value that is too small will result in too few points used for plane fitting, and the deformation at each point will be greater than that. d The system is insensitive to deformation detection; the larger the d-value, the slower the computation and the greater the similarity of the fitted planes from different points, resulting in insensitivity to small distances; for points... i ( x i , y i , z i ), for its corresponding set D i Elements p ij and d Values ​​are compared, if p ij ≤ d Then the point j ( x j , y j , z j The coordinates of ) are placed in the set K i ,in i =1…… n ; Step 3.3: Determine the best-fit plane using the SVD method: Compute set K i Midpoint cloud coordinates X , Y , Z their respective means X mean ,Y mean ,Z mean Using sets K i Construct a coefficient matrix from all elements: ； For matrix A Perform singular value decomposition. A = UDV T ,in: D It is a diagonal matrix. U and V All are unitary matrices; the eigenvectors corresponding to the smallest singular values ​​are the coefficient vectors of the fitting plane. coefficient ; ; ; The specific calculation process in step 4 is as follows: Calculation points i to its best-fit plane distance : ； The specific calculation process in step 5 is as follows: Initial point of 3D laser scanning and points ( Coordinates, substitute point Corresponding equation )*( If the result is negative, the distance is negative; otherwise, it is positive and corrected.

2. The method for describing rock mass deformation and deformation trend based on three-dimensional laser scanning point cloud data according to claim 1, characterized in that, The preprocessing of the original point cloud in step 2 includes deleting point clouds outside the study area, point cloud denoising, and resampling.

3. The method for describing rock mass deformation and deformation trend based on three-dimensional laser scanning point cloud data according to claim 1, characterized in that, The specific coloring process in step 6 is as follows: Calculate the distance between each point Assign each point in the point cloud and use RGB coloring based on the calculated distance to obtain a point-to-surface distance color map. In the map, points with closer calculated distances are distinguished by one color, and the area is indicated to be relatively flat, thus reducing attention. Positive and negative distance values ​​are represented by different colors.

4. The method for describing rock mass deformation and deformation trend based on three-dimensional laser scanning point cloud data according to claim 1, characterized in that, The specific operation process in step 7 is as follows: Based on the point-to-area distance color map, statistically analyze the distance distribution characteristic parameters in the brightly colored areas. , ; For the local maximum deformation within the region, This represents the average deformation level within the region.

Images