A coal mine tunnel deformation detection method and system based on point cloud

Abstract

The present application belongs to the field of intelligent detection of coal mine roadway, and provides a coal mine roadway deformation detection method and system based on point cloud, which is characterized in that the technical scheme is to perform feature registration, section registration and overall alignment of the roadway on the reference and target three-dimensional point cloud data; project the registered reference three-dimensional point cloud data and target three-dimensional point cloud data to a two-dimensional plane to obtain a two-dimensional scatter plot, extract the boundary points of the roadway based on the two-dimensional scatter plot, fit the boundary line of the entire roadway based on the boundary points of the roadway, calculate the trend curve of the roadway, move the cross section along the trend curve of the roadway at equal intervals, extract the point cloud section profile corresponding to each position and calculate the local width of the bounding box to obtain the minimum bounding box; emit a ray to the roadway based on the minimum bounding box, find the point on the registered target point cloud and the point or adjacent point on the reference point cloud data through which the ray passes, and calculate the distance between the two points, and calculate the deformation of the to-be-detected point based on the calculated distance. It can stably operate in a complex environment of the roadway.

Description

Technical Field

[0001] This invention belongs to the field of intelligent detection technology for coal mine roadways, and particularly relates to a method and system for detecting deformation in coal mine roadways based on point cloud. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Coal mine roadways are typically located in enclosed, humid, and high-risk underground environments, making their stability crucial for mine safety. Traditional deformation detection methods, including cross-sectional methods and roof delamination monitoring, rely on manual measurements, which suffer from problems such as scattered measurement points, large measurement errors, and long processing times. Furthermore, methods that compare design values ​​with measured values ​​are significantly affected by complex underground geological conditions, making it difficult to obtain accurate deformation information. In recent years, three-dimensional laser scanning technology has gradually become the main means of roadway deformation detection. However, while static three-dimensional laser scanning offers high accuracy, it is limited by the complex internal structure and uneven external contours of underground roadways, failing to meet the requirements for efficient detection. Summary of the Invention

[0004] To address at least one of the technical problems mentioned above, this invention provides a point cloud-based method and system for detecting deformation in coal mine roadways. It utilizes lidar to acquire complete point cloud information of the roadway, first generating a projection onto the roadway through an external minimum bounding box, and then calculating horizontal and vertical deformations through overall and local profiles to analyze the deformation. This method offers higher measurement efficiency and accuracy, and can operate stably in complex roadway environments. It can more accurately reflect overall and local changes in the roadway during deformation detection, meeting the actual needs of coal mine production.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A first aspect of the present invention provides a method for detecting deformation in coal mine roadways based on point clouds, comprising the following steps:

[0007] The acquired baseline 3D point cloud data and target 3D point cloud data are subjected to feature registration, cross-section registration and overall alignment of the roadway to obtain the registered baseline 3D point cloud data and target 3D point cloud data;

[0008] The registered 3D point cloud data are projected onto a 2D plane to obtain a 2D scatter plot. The boundary points of the tunnel are extracted based on the 2D scatter plot. The boundary line of the entire tunnel is fitted based on the boundary points of the tunnel. The tunnel trend curve is calculated. The cross section is moved at equal intervals along the tunnel trend curve. At each position, the point cloud contour of the cross section is extracted based on the target point cloud and the local width of the bounding box is calculated to generate the minimum bounding box that matches the geometry of the tunnel.

[0009] A ray is emitted into the tunnel based on the minimum bounding box. The ray is used to find points on the registered target point cloud and points on the reference point cloud data or nearby points, and the distance between the two points is calculated. The deformation of the point to be detected is calculated based on the calculated distance.

[0010] Furthermore, when performing feature registration on the acquired reference 3D point cloud data and target 3D point cloud data, feature vectors of the reference 3D point cloud data and target 3D point cloud data are extracted. The first feature vector is the direction with the largest variance, representing the extension direction of the roadway point cloud. x The axis direction, the second eigenvector is the direction with the second largest variance. y The direction of the axis, the direction of the third eigenvector with minimum variance represents z Axial direction.

[0011] Furthermore, when performing cross-sectional registration between the reference 3D point cloud data and the target 3D point cloud data, the least squares method is used to fit the centerline of the roadway, and cross-sectional registration is performed.

[0012] Furthermore, the extraction of roadway boundary points based on a two-dimensional scatter plot includes:

[0013] Cut multiple cross sections perpendicular to the direction of the tunnel extension;

[0014] The target point cloud data is projected onto each cross-sectional plane to identify the boundaries of the left and right sides and the top and bottom plates. Among them, based on the top view of the sliced ​​point cloud, the boundary lines of the left and right sides are extracted by analyzing the spatial distribution characteristics of the point cloud. The point cloud density at each height along the longitudinal direction is counted, and the two height values ​​with the densest point cloud distribution are selected as the top plate line and bottom plate line of the roadway cross section, respectively.

[0015] Furthermore, the process of fitting the boundary line of the entire roadway based on the boundary points of the roadway and calculating the roadway trend curve includes:

[0016] Take the midpoint of the left and right boundaries of each cross section as the center point of that cross section, and use this point as the fitting control point for each cross section. Sequence of all cross-sectional center points These are considered as control points for curve fitting and are assigned weights. Control points and their corresponding weights based on curve fitting The boundary line of the entire tunnel is fitted using a B-spline function to calculate the tunnel trend curve.

[0017] Furthermore, the fitting function for the boundary line of the entire tunnel using B-spline functions is:

[0018] ,

[0019] ,

[0020] ,

[0021] in, The curve represents the trend of the tunnel. For each cross-section, the fitting control points, Weights are assigned to the fitting control points for each cross-section. For B-spline basis functions, For parameters, It is a zero-degree basis function. for i +1 node k -1 degree basis functions For the first i Each node value For the first i + k -1 node value, for i At the node k -1 degree basis functions For the first i + k Each node value.

[0022] Furthermore, the step of emitting rays into the lane based on the minimum bounding box and searching for points on the registered target point cloud and points or neighboring points on the reference point cloud data through which the rays pass includes:

[0023] The smallest bounding box emits rays into the tunnel for the reference point cloud. P and target point cloud Q When the ray on the bounding box passes through the point on the target point cloud At that time, record the coordinates of the point and continue emitting rays; at the reference point cloud P If the ray hits one of the points directly, then record the coordinates of that point. If the point has not been visited, the K-nearest neighbor algorithm is used to select the nearest point. Calculation points Coordinates of the corresponding point ,in k The number of nearest neighbor points selected for the target point cloud. , and Let each represent the first point in the target point cloud. i The three-dimensional coordinates of the neighboring points; For a corresponding point in the reference point cloud 3D coordinates; , and Represents a point in the target point cloud qPosition in a three-dimensional coordinate system , and Represents the corresponding point in the reference point cloud. The three-dimensional coordinates.

[0024] A second aspect of the present invention provides a point cloud-based coal mine roadway deformation detection system, comprising:

[0025] The registration module is used to perform feature registration, cross-section registration, and overall alignment of the acquired reference 3D point cloud data and target 3D point cloud data to obtain the registered reference 3D point cloud data and target 3D point cloud data.

[0026] The bounding box generation module is used to project the registered 3D point cloud data onto a 2D plane to obtain a 2D scatter plot, extract the boundary points of the tunnel based on the 2D scatter plot, fit the boundary line of the entire tunnel based on the boundary points of the tunnel, calculate the tunnel trend curve, move the cross section at equal intervals along the tunnel trend curve, extract the point cloud contour of the cross section based on the target point cloud at each position and calculate the local width of the bounding box, and generate the minimum bounding box that matches the geometry of the tunnel.

[0027] The deformation detection module is used to emit rays into the roadway based on the minimum bounding box, find points on the registered target point cloud and points on the reference point cloud data or neighboring points through which the rays pass, calculate the distance between the two points, and calculate the deformation of the point to be detected based on the calculated distance.

[0028] A third aspect of the present invention provides a computer-readable storage medium.

[0029] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the point cloud-based coal mine roadway deformation detection method described above.

[0030] A fourth aspect of the present invention provides a computer device.

[0031] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps of the point cloud-based coal mine roadway deformation detection method described above.

[0032] Compared with the prior art, the beneficial effects of the present invention are:

[0033] 1. This invention is based on mobile three-dimensional laser scanning technology, which reduces measurement time and improves detection efficiency compared with traditional manual inspection. At the same time, it can operate stably in narrow, low-light, and high-dust environments such as coal mine roadways, meeting the needs of practical applications.

[0034] 2. This invention utilizes B-spline interpolation technology to accurately obtain the center curve of the tunnel extension and generates a minimum bounding box by moving the cross-section along the curve, effectively reducing the computational complexity of the bounding box. This allows cloud feature extraction at different time points to be based on a unified curve reference, improving the robustness and stability of feature extraction and avoiding errors caused by differences in viewpoint and pose in traditional global bounding box methods.

[0035] 3. This invention uses ray emission from the surface of a minimum bounding box and Euclidean distance to calculate the displacement between corresponding points in the point cloud. This suppresses the interference of complex structures and noise within the tunnel on deformation calculation, thereby improving the overall analytical capability of deformation detection. Simultaneously, ray detection focuses only on data in the boundary region, significantly reducing the number of invalid point matches, lowering the computational burden, and ensuring computational efficiency.

[0036] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0037] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0038] Figure 1 This is a flowchart of a point cloud-based method for detecting deformation in coal mine roadways, provided by an embodiment of the present invention.

[0039] Figure 2 This is a 3D point cloud mapping to a 2D scatter plot provided in an embodiment of the present invention; wherein, (a) represents 3D point cloud data, (b) represents a scatter plot of plane 1 mapped to a cross section, and (c) represents a scatter plot of plane 2 mapped to a horizontal plane;

[0040] Figure 3 This is a schematic diagram of the fitting curve provided in an embodiment of the present invention;

[0041] Figure 4 This is a schematic diagram of tunnel cross-sectional deformation detection based on the ray method provided in an embodiment of the present invention;

[0042] Figure 5 This is a schematic diagram of a point cloud-based method for detecting deformation in coal mine roadways, provided in an embodiment of the present invention. Detailed Implementation

[0043] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0044] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0045] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0046] Example 1

[0047] like Figure 1 As shown in the figure, this embodiment provides a method for detecting deformation in coal mine roadways based on point clouds, including the following steps:

[0048] Step 1: Perform feature registration and cross-sectional registration on the acquired reference 3D point cloud data and target 3D point cloud data to obtain registered reference point cloud data and target point cloud data;

[0049] Specifically, the steps include the following:

[0050] Step 101: Process the reference point cloud data P and target point cloud data Q Perform coarse registration to align the main directions;

[0051] In this embodiment, principal component analysis is used to perform coarse registration of the point cloud, which specifically includes the following steps:

[0052] Step 1011: Extract the feature vectors from the original point cloud data and the point cloud data to be registered, respectively;

[0053] In this embodiment, the reference point cloud data P and target point cloud data Q The eigenvectors are in a three-dimensional coordinate system O-xyz The components are orthogonal in three directions, and the three axes satisfy the right-hand rule. Among them, the 𝑥 axis is defined as the main direction of the longitudinal extension of the tunnel, corresponding to the longest direction of the tunnel; the 𝑦 axis is defined as the width direction of the tunnel, used to describe the spatial span between the two sides; and the 𝑧 axis is defined as the height direction of the tunnel, pointing towards the roof.

[0054] Feature vectors are extracted from the baseline 3D point cloud data and the target 3D point cloud data. The first feature vector, representing the direction with the largest variance, indicates the extension direction of the tunnel point cloud. x The axis direction, the second eigenvector is the direction with the second largest variance. yThe direction of the axis, the direction of the third eigenvector with minimum variance represents z Axial direction.

[0055] Step 1012: Calculate the rotation and translation matrices, and apply the rotation and translation matrices to the point cloud to be registered to perform a rigid transformation, so that the target point cloud data... Q Feature vectors and baseline point cloud data P Feature vector alignment is used to achieve coarse registration of point cloud data from two periods.

[0056] In this embodiment, the optimal rotation and translation matrix is ​​calculated in the XOY plane by combining the ICP concept with the maximum correlation coefficient method.

[0057] Specifically, the baseline point cloud data is first calculated. P and target point cloud data Q center of mass and , is represented as:

[0058] , ,in, For the i-th point in the reference point cloud, For the j-th point in the target point cloud, The number of points.

[0059] Subsequently, benchmark point cloud data was constructed. P and target point cloud data Q The covariance matrix between ;

[0060] Finally, singular value decomposition is performed on matrix H to obtain... Then the rotation matrix and translation matrix can be calculated using the following formula: ,in, From target point cloud data Q An orthogonal matrix formed by the eigenvectors of , Baseline point cloud data P An orthogonal matrix composed of eigenvectors;

[0061] The minimum rotation angle is set for each iteration. The variance of the distance between the nearest points in two sets of point clouds is calculated continuously. The formula for calculating the variance is as follows: ,in x Let be the distance from each point in the point cloud to the nearest point in the target point cloud. for x expectation, when the variance The optimal rotation angle is obtained when the minimum value is reached.

[0062] Step 102: Perform cross-sectional registration on the point cloud data after coarse registration;

[0063] In this embodiment, the least squares method is used to fit the centerline of the tunnel in the XOZ plane, and cross-section registration is performed.

[0064] Specifically, the steps include the following:

[0065] Step 1021: Rotate the normal vector of the base plane of the point cloud to be registered to the normal vector of the base plane of the target point cloud, with the rotation center located at the center of the minimum bounding box of the point cloud.

[0066] A subset is randomly selected from the sample, and the least squares estimation algorithm is used to calculate the parameters of the tunnel point cloud model.

[0067] Step 1022: Calculate the deviation of all samples in the model and compare it with a pre-set threshold. When the deviation is less than the threshold, mark the sample point as an inside line; otherwise, mark it as an outlier.

[0068] Step 1023: Record the current number of inner line points, and repeat steps 1021-1023;

[0069] Step 1024: In each iteration, record the best model parameters and display the most interior points;

[0070] Step 1025: Calculate the iteration termination criterion based on the expected error rate, the optimal number of inner layers, the total number of samples, and the current iteration count. Decide whether to terminate the iteration based on the termination condition. If the iteration terminates, use the best model parameters as the final estimated model parameters.

[0071] Step 103: Based on the point cloud data after the tunnel cross-section registration, perform overall alignment of the tunnel;

[0072] In this embodiment, in the YOZ plane, by fitting the tunnel floor and rotating and translating it, the entire system is horizontally aligned, thus achieving accurate registration of the point cloud.

[0073] Step 1031: Using the idea of ​​the nearest point iterative algorithm, combined with the point cloud correlation coefficient, calculate the original point cloud data. The rotation and translation matrices between and the point cloud to be registered ;

[0074] Step 1032: Subsequently, the nearest point iterative algorithm is used to minimize the difference between the two point clouds. In each iteration, the algorithm selects the nearest point as the corresponding point and calculates the transformation. Minimize the equation:

[0075] ,

[0076] in, Let be the error function. and Point clouds P and Q The number of points to be registered in the middle The weights for point matching. For points in the reference point cloud, For the points in the target point cloud, Let be a rotation matrix. It is a translation matrix;

[0077] By following the steps above, the original point cloud data and the 3D point cloud data to be registered can be aligned as a whole.

[0078] Step 2: Project the registered 3D point cloud data onto a 2D plane to obtain a 2D scatter plot. Extract the boundary points of the tunnel based on the 2D scatter plot. Fit the boundary line of the entire tunnel based on the boundary points of the tunnel. Calculate the tunnel trend curve. Move the cross section at equal intervals along the tunnel trend curve. Extract the point cloud contour of the cross section based on the target point cloud at each position and calculate the local width of the bounding box to generate the minimum bounding box that matches the tunnel geometry.

[0079] Specifically, the steps include the following:

[0080] Step 201: Slice the registered 3D point cloud data, project the sliced ​​point cloud data onto a 2D plane to obtain a 2D scatter plot, and extract the boundary points of the tunnel based on the 2D scatter plot;

[0081] In this embodiment, slicing the registered 3D point cloud data specifically includes:

[0082] The main direction of the tunnel is extracted based on the registered 3D point cloud data. At the same time, the cross-sectional point cloud perpendicular to the tunnel extension direction is extracted as a tunnel slice.

[0083] In this embodiment, the sliced ​​point cloud data is projected onto a two-dimensional plane to obtain a two-dimensional scatter plot, specifically including:

[0084] Projecting the point cloud of the tunnel slices onto... plane and On a plane, the cross-sectional view and the plan view of the slice are shown respectively, as follows: Figure 2 As shown, the dashed lines represent slices of the point cloud of a rectangular tunnel, perpendicular to the direction of tunnel extension. Figure 2 In the middle (a), the data represents three-dimensional point cloud data. Figure 2 (b) shows a scatter plot of plane 1 mapped to the cross section. Figure 2 In the middle (b), a scatter plot is shown on the plane 2 mapped to the horizontal plane.

[0085] In this embodiment, the boundary points of the tunnel are extracted based on a two-dimensional scatter plot, and then analyzed. Figure 2The distribution density of two-dimensional scattered points along the directions shown in plane 1 and plane 2 is used to extract the boundary points of the tunnel, specifically including:

[0086] First, take multiple cross-sections perpendicular to the direction of the tunnel's extension;

[0087] Specifically, in x Point cloud fragments on each cross-section are extracted by interpolating planes at certain intervals along the axial direction. For the point set of each cross-section, kernel density analysis and edge detection algorithms are used to extract the boundary points of the tunnel contour, such as the boundaries of the left and right side plates.

[0088] Then the target point cloud data is projected onto each cross-sectional plane to identify the boundaries of the left and right sidewalls and the top and bottom plates;

[0089] Specifically, based on the top view of the sliced ​​point cloud, the spatial distribution characteristics of the point cloud are analyzed to extract the boundary lines on the left and right sides, thereby determining the spatial location of the two sides of the tunnel.

[0090] The point cloud density at each height along the longitudinal direction is statistically analyzed, and the two values ​​with the densest point cloud distribution are selected as the top and bottom lines of the roadway cross-section, respectively.

[0091] Step 202: Fit the boundary line of the entire tunnel based on the boundary points of the tunnel, and calculate the tunnel trend curve;

[0092] First, take the midpoint of the left and right boundaries of each cross section as the center point of that cross section, and use this point as the fitting control point for each cross section. Sequence of all cross-sectional center points These are considered as control points for curve fitting and are assigned weights. The weight of a control point represents the degree of attachment of each control point;

[0093] Next, based on the control points and their corresponding weights, the boundary line of the entire tunnel is fitted using a B-spline function to calculate the tunnel trend curve.

[0094] Specifically, identifying basis functions nodes u The parameter space is divided into n There are nodes, ensuring that the values ​​within each node are monotonically increasing, i.e., extending along the alleyway. x The curve advances smoothly and continuously in the direction of travel.

[0095] The fitting formula is:

[0096] ,

[0097] ,

[0098] ,

[0099] in, The curve represents the trend of the tunnel. For each cross-section, the fitting control points, Weights are assigned to the fitting control points for each cross-section. For B-spline basis functions, For parameters, It is a zero-degree basis function. for i +1 node k -1 degree basis functions For the first i Each node value For the first i + k -1 node value, for i At the node k -1 degree basis functions For the first i + k Each node value;

[0100] like Figure 3 The image shows the fitted roadway trend curve, where circular dots represent curve control points, triangular dots represent K-value points, and the black curve represents the fitted curve. Utilizing the smoothing properties of B-spline curves, the roadway's curve direction can be accurately represented.

[0101] Step 203: Then move the cross section at equal intervals along the roadway curve, extract the point cloud cross section profile corresponding to the sampling center point of each cross section slice, and calculate the local width of the bounding box, finally forming the minimum bounding box that highly matches the geometry of the roadway.

[0102] Specifically, for each slice, the point cloud is projected onto the cross-sectional plane, the boundaries of the left and right sides and the top and bottom plates are identified, and the maximum lateral span is calculated to obtain the local width value. By smoothly fitting the local width of all sections, a continuous width function is generated. Then, combined with the tunnel trend curve, the bounding rectangle is generated segment by segment and connected along the depth direction to finally form a global minimum bounding box that highly matches the actual geometry of the tunnel.

[0103] This invention solves the bounding box by sliding the cross section along the tunnel axis, which effectively reduces the global computational complexity and avoids the data redundancy caused by the previous bounding box generation based on fixed coordinate axes. At the same time, the same curve parameters and movement strategy are used for scanning point clouds at different times, so that the features of point clouds at different times have a consistent extraction benchmark, which significantly improves the robustness and stability of feature extraction.

[0104] Step 3: Based on the minimum bounding box, emit rays into the tunnel, find the points on the registered target point cloud and the reference point cloud data or neighboring points through which the rays pass, and calculate the distance between the two points. Calculate the deformation of the point to be detected based on the calculated distance.

[0105] In this embodiment, the ray casting method is used to search for neighboring points around the point of interest in the two-phase point clouds after registration, and the deformation value of the roadway is detected by calculating the distance between the midpoint of the two-phase point clouds and its corresponding point.

[0106] Figure 4 This is a schematic diagram of tunnel cross-sectional deformation detection based on the ray method, which specifically includes the following steps:

[0107] Step 301: Based on the minimum bounding box, emit a ray into the alleyway, find the points on the registered target point cloud and the reference point cloud data or neighboring points through which the ray passes, and calculate the distance between the two points;

[0108] In this embodiment, the minimum bounding box emits rays into the tunnel, for the reference point cloud. P and target point cloud Q When the ray on the bounding box passes through the point on the target point cloud At that time, record the coordinates of the point and continue emitting rays;

[0109] Secondly, in the benchmark point cloud P If the ray hits one of the points directly, then record the coordinates of that point. ,in, For a corresponding point in the reference point cloud 3D coordinates;

[0110] If the point has not been visited, the K-nearest neighbor algorithm is used to select the nearest point. Calculation points Coordinates of the corresponding point ,in k The number of nearest neighbor points selected for the target point cloud. , and Let each represent the first point in the target point cloud. i The three-dimensional coordinates of the neighboring points;

[0111] Finally, calculate the midpoint of the target point cloud. q Its corresponding point The distance between them is , indicating a point q The amount of deformation, , and Represents a point in the target point cloud q Position in a three-dimensional coordinate system , and Represents the corresponding point in the reference point cloud. 3D coordinates;

[0112] The accuracy of this method depends on the density of the point cloud dataset and the number of nearest neighbors. When the point cloud resolution is low and the number of nearest neighbors is small, the accuracy of the detection algorithm may be affected.

[0113] Step 302: Compare the distances from the point to be detected to the center line of the corresponding point in the target point cloud. Calculate the deformation of the point to be detected;

[0114] When calculating distances, rock stress and coal mine roadway deformation are taken into account, and the horizontal and vertical changes between corresponding points are calculated separately using distance.

[0115] like Figure 5 As shown, the distance from each point on the tunnel cross-section to the center line is the distance from that point to the center line. Figure 5 The magnified portion represents the distance difference between corresponding points:

[0116] This indicates the deformation at that point. express x Difference in direction, express y Difference in direction;

[0117] When phenomena such as roof collapse, sidewall bulging, and floor heaving occur in the tunnel, .

[0118] At the same time, utilizing and This allows for the further restoration of the horizontal and vertical deformations of the roadway at this cross-section to the original point cloud, and visualization of the deformation location and deformation value.

[0119] The method of this invention can achieve high-precision deformation detection of roadway point clouds, meeting the needs of coal mine safety monitoring.

[0120] Example 2

[0121] This embodiment provides a point cloud-based coal mine roadway deformation detection system, including:

[0122] The registration module is used to perform feature registration, cross-section registration, and overall alignment of the acquired reference 3D point cloud data and target 3D point cloud data to obtain the registered reference 3D point cloud data and target 3D point cloud data.

[0123] The bounding box generation module is used to project the registered 3D point cloud data onto a 2D plane to obtain a 2D scatter plot, extract the boundary points of the tunnel based on the 2D scatter plot, fit the boundary line of the entire tunnel based on the boundary points of the tunnel, calculate the tunnel trend curve, move the cross section at equal intervals along the tunnel trend curve, extract the point cloud contour of the cross section based on the target point cloud at each position and calculate the local width of the bounding box, and generate the minimum bounding box that matches the geometry of the tunnel.

[0124] The deformation detection module is used to emit rays into the roadway based on the minimum bounding box, find points on the registered target point cloud and points on the reference point cloud data or neighboring points through which the rays pass, calculate the distance between the two points, and calculate the deformation of the point to be detected based on the calculated distance.

[0125] It should be noted that the specific implementation of the point cloud-based coal mine roadway deformation detection system in this embodiment of the invention is similar to the specific implementation of the point cloud-based coal mine roadway deformation detection method in this embodiment of the invention. Please refer to the description in the method section for details. To reduce redundancy, it will not be repeated here.

[0126] Example 3

[0127] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the point cloud-based coal mine roadway deformation detection method described above.

[0128] Example 4

[0129] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the point cloud-based coal mine roadway deformation detection method described above.

[0130] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0131] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0132] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0133] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0134] 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.

[0135] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for detecting deformation in coal mine roadways based on point clouds, characterized in that, Includes the following steps: The acquired baseline 3D point cloud data and target 3D point cloud data are subjected to feature registration, cross-section registration and overall alignment of the roadway to obtain the registered baseline 3D point cloud data and target 3D point cloud data; The registered 3D point cloud data are projected onto a 2D plane to obtain a 2D scatter plot. Boundary points of the tunnel are extracted based on the 2D scatter plot. The boundary line of the entire tunnel is fitted based on the boundary points of the tunnel. The tunnel trend curve is calculated. The cross section is moved at equal intervals along the tunnel trend curve. At each position, the point cloud contour of the cross section is extracted based on the target point cloud and the local width of the bounding box is calculated. The minimum bounding box that matches the tunnel geometry is generated. For each slice, the point cloud is projected onto the cross section plane, the boundaries of the left and right sides and the top and bottom plates are identified, and the maximum lateral span is calculated to obtain the local width value. By smoothing the local width of all cross sections, a continuous width function is generated. Then, combined with the tunnel trend curve, the bounding rectangle is generated segment by segment and connected along the depth direction to finally form a global minimum bounding box that highly matches the actual geometry of the tunnel. Ray detection is based on the minimum bounding box. Rays are emitted into the roadway to find points on the registered target point cloud and points on the reference point cloud data or nearby points through which the rays pass. The distance between the two points is calculated, and the deformation of the point to be detected is calculated based on the calculated distance. This suppresses the interference of complex structures and noise inside the roadway on deformation calculation, thereby improving the overall deformation detection resolution. Ray detection only focuses on the data in the boundary area, which greatly reduces the number of invalid point matches, reduces the computational burden, and ensures computational efficiency. The step of emitting rays into the roadway based on the minimum bounding box and finding points or neighboring points on the registered target point cloud and the reference point cloud data through which the rays pass includes: The smallest bounding box emits rays into the tunnel for the reference point cloud. P and target point cloud Q When the ray on the bounding box passes through the point on the target point cloud At that time, record the coordinates of the point and continue emitting rays; at the reference point cloud P If the ray hits one of the points directly, then record the coordinates of that point. If the point has not been visited, the K-nearest neighbor algorithm is used to select the nearest point. Calculation points Coordinates of the corresponding point ,in k The number of nearest neighbor points selected for the target point cloud. , and Let each represent the first point in the target point cloud. i The three-dimensional coordinates of the neighboring points; For a corresponding point in the reference point cloud 3D coordinates; , and Represents a point in the target point cloud q Position in a three-dimensional coordinate system , and Represents the corresponding point in the reference point cloud. The three-dimensional coordinates.

2. The method for detecting deformation in coal mine roadways based on point clouds as described in claim 1, characterized in that, When performing feature registration on the acquired reference 3D point cloud data and target 3D point cloud data, feature vectors of the reference 3D point cloud data and target 3D point cloud data are extracted. The first feature vector is the direction with the largest variance, representing the extension direction of the tunnel point cloud. x The axis direction, the second eigenvector is the direction with the second largest variance. y The direction of the axis, the direction of the third eigenvector with minimum variance represents z Axial direction.

3. The method for detecting deformation in coal mine roadways based on point clouds as described in claim 1, characterized in that, When performing cross-sectional registration between the reference 3D point cloud data and the target 3D point cloud data, the least squares method is used to fit the centerline of the tunnel, and cross-sectional registration is performed.

4. The method for detecting deformation in coal mine roadways based on point clouds as described in claim 1, characterized in that, The extraction of roadway boundary points based on a two-dimensional scatter plot includes: Cut multiple cross sections perpendicular to the direction of the tunnel extension; The target point cloud data is projected onto each cross-sectional plane to identify the boundaries of the left and right sides and the top and bottom plates. Specifically, based on the top view of the sliced ​​point cloud, the spatial distribution characteristics of the point cloud are analyzed to extract the boundary lines on the left and right sides, and statistical analysis is performed along the longitudinal direction. Point cloud density at altitude: select the two points with the densest point cloud distribution. The values ​​are used as the top and bottom lines of the roadway cross-section, respectively.

5. The method for detecting deformation in coal mine roadways based on point clouds as described in claim 1, characterized in that, The process of fitting the boundary line of the entire tunnel based on the boundary points of the tunnel and calculating the tunnel trend curve includes: Take the midpoint of the left and right boundaries of each cross section as the center point of that cross section, and use this point as the fitting control point for each cross section. Sequence of all cross-sectional center points These are considered as control points for curve fitting and are assigned weights. Control points and their corresponding weights based on curve fitting The boundary line of the entire tunnel is fitted using a B-spline function to calculate the tunnel trend curve.

6. The method for detecting deformation in coal mine roadways based on point clouds as described in claim 1, characterized in that, The fitting function for the boundary line of the entire tunnel using B-spline functions is: ， ， ， in, The curve represents the trend of the tunnel. For each cross-section, the fitting control points, Weights are assigned to the fitting control points for each cross-section. For B-spline basis functions, For parameters, It is a zero-degree basis function. for i +1 node Secondary basis functions For the first i Each node value For the first i + k -1 node value, for i At the node Secondary basis functions For the first i + k Each node value.

7. A point cloud-based coal mine roadway deformation detection system, characterized in that, include: The registration module is used to perform feature registration, cross-section registration, and overall alignment of the acquired reference 3D point cloud data and target 3D point cloud data to obtain the registered reference 3D point cloud data and target 3D point cloud data. The bounding box generation module projects the registered 3D point cloud data onto a 2D plane to obtain a 2D scatter plot. Based on the 2D scatter plot, it extracts the boundary points of the tunnel, fits the boundary line of the entire tunnel based on the boundary points, calculates the tunnel trend curve, moves the cross section at equal intervals along the tunnel trend curve, extracts the point cloud contour of the cross section based on the target point cloud at each position, calculates the local width of the bounding box, and generates a minimum bounding box that matches the tunnel geometry. For each slice, the point cloud is projected onto the cross section plane, the boundaries of the left and right sides and the top and bottom plates are identified, and the maximum lateral span is calculated to obtain the local width value. By smoothly fitting the local width of all cross sections, a continuous width function is generated. Then, combined with the tunnel trend curve, the bounding rectangle is generated segment by segment and connected along the depth direction to finally form a global minimum bounding box that highly matches the actual geometry of the tunnel. The deformation detection module is used to emit rays into the roadway based on the minimum bounding box, find points on the registered target point cloud and points on the reference point cloud data or neighboring points through which the rays pass, and calculate the distance between the two points. Based on the calculated distance, the deformation of the point to be detected is calculated. It suppresses the interference of complex structures and noise inside the roadway on deformation calculation, thereby improving the overall deformation detection resolution. Ray detection only focuses on the data in the boundary area, which greatly reduces the number of invalid point matching, reduces the computational burden, and ensures computational efficiency. The step of emitting rays into the roadway based on the minimum bounding box and finding points or neighboring points on the registered target point cloud and the reference point cloud data through which the rays pass includes: The smallest bounding box emits rays into the tunnel for the reference point cloud. P and target point cloud Q When the ray on the bounding box passes through the point on the target point cloud At that time, record the coordinates of the point and continue emitting rays; at the reference point cloud P If the ray hits one of the points directly, then record the coordinates of that point. If the point has not been visited, the K-nearest neighbor algorithm is used to select the nearest point. Calculation points Coordinates of the corresponding point ,in k The number of nearest neighbor points selected for the target point cloud. , and Let each represent the first point in the target point cloud. i The three-dimensional coordinates of the neighboring points; For a corresponding point in the reference point cloud 3D coordinates; , and Represents a point in the target point cloud q Position in a three-dimensional coordinate system , and Represents the corresponding point in the reference point cloud. The three-dimensional coordinates.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the point cloud-based coal mine roadway deformation detection method as described in any one of claims 1-6.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the point cloud-based coal mine roadway deformation detection method as described in any one of claims 1-6.

Images