Construction method and system for transparent working face of coal mine geology

Abstract

The invention discloses a construction method and system for a coal mine geological transparent working face, and relates to the technical field of intelligent mine and geological information, and the method comprises the following steps: collecting and preprocessing visible light, near-infrared images and point cloud data, extracting texture, reflectivity, curvature and density features, and calculating the texture, reflectivity, curvature and density features; the method comprises the following steps: constructing a three-dimensional structure reconstruction model, a rock stratum physical property inversion model and a geological stress prediction model, generating a three-dimensional geometric structure coordinate set, a rock stratum moisture content matrix and a vertical stress distribution matrix of each region, setting an anomaly judgment condition, outputting three-dimensional coordinates of an abnormal region, carrying out graded early warning, synthesizing a three-dimensional visual image, and predicting the geological stress. And holographic transparent expression of the working face geological state is realized. According to the method, centimeter-level three-dimensional reconstruction of the geological transparent working face of the coal mine, precise positioning of a hidden water-containing area and a stress abnormal area and construction of a full-attribute dynamic visual scene are achieved, the timeliness and accuracy of early warning of disasters such as water inrush and rockburst are remarkably improved, real-time geological decision support is provided for intelligent mining, and safe and efficient production is guaranteed.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent mines and geological information technologies, and particularly relates to a method and system for constructing a transparent coal mine geological working face. Background Art

[0002] Coal mine geology is a geological discipline that studies the formation, distribution, and mining conditions of coal deposits, mainly involving the occurrence law of coal seams, the characteristics of surrounding rocks, tectonic features, and hydrogeological conditions. By analyzing the stratigraphic structure, the physical and mechanical properties of coal and rock, the development degree of faults and folds, and the groundwater migration law, it evaluates the coal mine resource reserves, mining feasibility, and potential disaster risks. In the coal mine exploration stage, coal mine geology determines the coal seam thickness, dip angle, and stability. During the mining process, it monitors the stability of the roof and floor, the occurrence of gas, and the water inrush channels, guiding the roadway layout and support design. At the same time, coal mine geology combines geophysical exploration, borehole data, and three-dimensional modeling technology to construct a mine geological model, providing a scientific basis for gas drainage, water hazard prevention, and rock burst early warning. As a bridge connecting resource development and safety production, coal mine geology supports the intelligent mining system to achieve precise geological navigation and risk prevention and control by revealing the interaction law between geological bodies and mining activities, and is a basic support discipline for the sustainable development of the modern coal industry.

[0003] To solve the problems of opaque geological information and lagging disaster early warning in coal mine mining, the existing technology is to use a method of combining single borehole sampling with static geological modeling for processing, but there will still be situations of low data coverage density and difficulty in fusing multi-source heterogeneous data, which in turn leads to a high false negative rate in the identification of hidden disaster-causing bodies and poor timeliness of model updates. To solve the above limitations, a method and system for constructing a transparent coal mine geological working face are proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for constructing a transparent coal mine geological working face to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: In the first aspect, a method for constructing a transparent coal mine geological working face includes the following steps: S 1, Collect and preprocess the visible light and near-infrared image data and point cloud coordinate data of the transparent coal mine geological working face; S 2, Extract the features of the transparent coal mine geological working face from the preprocessed visible light and near-infrared image data and point cloud coordinate data; S 3, Combine the features of the transparent coal mine geological working face to construct a three-dimensional structure reconstruction model, a rock formation physical property inversion model, and a geological stress prediction model, and output the three-dimensional geometric structure coordinate set of the transparent coal mine geological working face, the matrix of water content of each area of the rock formation, and the vertical stress distribution matrix; S4. Based on the output 3D geometric structure coordinate set, the rock moisture matrix of each region, and the vertical stress distribution matrix, design abnormal area determination conditions, output the abnormal volume, and issue graded warnings; S5. Synthesize a 3D visualization image showing the 3D geometric structure, the water content distribution of the rock layers in each region, and the vertical stress distribution.

[0006] A further improvement of the technical solution of the present invention is that in S1, the process of collecting and preprocessing the visible light and near-infrared image data and point cloud coordinate data of the transparent working surface of the coal mine geology includes: Explosion-proof multispectral imaging equipment is deployed at 20-meter intervals on the roof of the coal mine tunnel. The multispectral imaging equipment has a built-in spectroscopic prism component that simultaneously collects images in the 400-700nm visible light band and the 700-2500nm near-infrared band. The multispectral imaging equipment switches bands using a filter wheel, dynamically adjusts the exposure time of each band based on the underground light intensity, and outputs visible light and near-infrared image data. Phase-shifted 3D laser scanning equipment is staggered on both sides of the coal mine tunnel. The 3D laser scanning equipment emits near-infrared lasers with a modulation frequency of 150MHz, obtains the distance to the target point through the phase difference ranging principle, and simultaneously records the horizontal angle and pitch angle of the scan, and outputs point cloud coordinate data; The multispectral imaging equipment is installed 2.5 meters from the base plate, with a pitch angle of 30°. The scanning head of the 3D laser scanning equipment is 1.5 meters from the base plate, with a horizontal scanning angle of 120° and a vertical scanning angle of -30° to +45°. The 3D laser scanning equipment is connected by a time synchronization signal line. The raw visible and near-infrared image data were corrected using dark current matrix and gain parameters. The image resolution after correction remained at 5472 × 3648 pixels and was stored in 16-bit RAW format. An adaptive voxel filtering algorithm was used to divide the point cloud space into a 3 mm × 3 mm × 3 mm voxel grid, and the centroid points with a density ≥ 2000 points / m2 within each voxel were retained.

[0007] A further improvement of the technical solution of the present invention is that in S2, the process of extracting features from the pre-processed visible light and near-infrared image data and point cloud coordinate data includes: Extracting texture uniformity features based on preprocessed visible light image data. The process of extracting texture uniformity features includes using a local binary pattern algorithm to divide the image into 8×8 pixel local windows, performing a grayscale comparison between the central pixel and the neighboring pixels in each window, generating a 256-dimensional LBP code value, and obtaining the variance of the LBP code value within the window. Continuous regions with a variance less than 0.1 are marked as a complete rock formation boundary coordinate set. Based on the preprocessed near-infrared image data, extract the near-infrared reflectance difference features. The process of extracting the near-infrared reflectance difference features includes extracting the reflectance values in the 850nm and 1650nm bands. According to the moisture content sensitivity characteristics of underground rocks, set the dual-band reflectance ratio threshold. If the reflectance ratio of the 850nm and 1650nm bands is greater than the dual-band reflectance ratio threshold, it is determined as a high moisture content abnormal area, and output the spatial coordinate matrix of the high moisture content abnormal area; Based on the preprocessed point cloud coordinate data, extract the surface curvature mutation features and density distribution features; The process of extracting the surface curvature mutation features includes searching for 50 neighborhood points within a radius of 3cm of each target point, obtaining the local Gaussian curvature based on the difference between the normal vectors of the target point and its neighborhood points, and marking the area where the local Gaussian curvature is greater than 0.1mm -1 as the surface curvature mutation area, and output the three-dimensional coordinate sequence of the potential fracture zone; The process of extracting the density distribution features includes dividing the point cloud space into 10cm³ cubic voxel grids, counting the number of point clouds in each voxel, and using the ratio of the number of point clouds in each voxel to the volume of the cubic voxel grid as its density. Screen the voxels with a density less than 1500 points / m 3 and determine them as loose and broken zones, and generate a density distribution heat map of the loose and broken zones.

[0008] A further improvement of the technical solution of the present invention is that in S3, the process of constructing a three-dimensional structure reconstruction model and outputting the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face includes: Taking the texture uniformity features, surface curvature mutation features, and the point cloud coordinate set with a filtered density ≥ 2000 points per square meter as input data; Using an improved iterative closest point algorithm to iteratively solve the optimal rigid body transformation matrix , and the iterative solution process includes initializing the rigid body transformation matrix and aligning based on the centroid of the point cloud. Search for the nearest neighbor points of each point in the target point cloud, and construct a comprehensive objective function including the basic point cloud registration term , the complete rock layer boundary constraint term and the fracture zone curvature constraint term . Among them, the basic point cloud registration term minimizes the source point cloud and the target point cloud For the rigid body transformation error, the complete rock formation boundary constraint term forces the complete rock formation boundary points to match the borehole data points, and the fracture zone curvature constraint term increases the registration weight of the points in the curvature mutation area. Key point pairs are screened according to the complete rock formation boundary constraint term and the fracture zone curvature constraint term, and the optimal rigid body transformation matrix is solved by singular value decomposition. The maximum number of iterations is set, and the iteration stops until the maximum number of iterations is reached, generating a three-dimensional structure reconstruction model. The calculation process is as follows: ; ; ; ; ; ; Among them, is the coordinate of the complete rock formation boundary point measured by the borehole, is the texture constraint weight calibrated by 50 groups of registration experiments with known complete rock formation boundaries. The boundary refers to the complete rock formation area screened by the texture uniformity feature, is the curvature constraint weight optimized based on the registration error reduction rate of the fracture zone. The fracture refers to the potential fracture area marked by the curvature mutation feature, is the rotation matrix obtained by SVD, and are the left and right singular vector matrices, both of which are orthogonal matrices, represents the translation vector, and are the weighted centroids of the source point cloud and the target point cloud; Fuse the multi-view registration point clouds into a point set in a unified coordinate system, construct a point cloud surface mesh based on the Delaunay triangulation algorithm, connect adjacent points to generate triangular patches, make the maximum side length of the triangular patches ≤ 5 cm, and extract the vertex coordinates of the mesh and the topological relationship of the triangular patches after removing the hanging triangular patches, and output the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face.

[0009] A further improvement of the technical solution of the present invention lies in: in S3, the process of constructing a rock formation physical property inversion model and outputting a matrix of water content of each area of the rock formation includes: Taking the spatial coordinate matrix of the high water content anomaly area in the near-infrared reflectance difference feature as input data, based on the extension of the Lambert-Beer law, establishing a non-linear relationship between the reflectance ratio R and the water content w of the rock formation, and generating a rock formation physical property inversion model. The calculation process is as follows: ; Adjust the coefficients of the rock stratum physical property inversion model through the Bayesian optimization algorithm, introduce the measured water content samples of the rock stratum in the borehole, and minimize the root mean square error between the predicted value of the rock stratum water content and the measured value of the rock stratum water content sample in the borehole; Align the spatial coordinate matrix of the high water content abnormal area with the three-dimensional grid of the coal mine geological transparent working face, calculate the reflectivity ratio grid by grid and invert the water content of the rock stratum. For the non-abnormal area of the rock stratum water content, use Kriging interpolation to fill it. Combine the measured rock stratum water content sample data in the borehole, calibrate the output of the rock stratum physical property inversion model through weighted least squares method, generate the rock stratum water content matrix of each area of the coal mine geological transparent working face, store it as floating-point raster data, set the rock stratum water content threshold, and determine the area where the predicted value of the rock stratum water content exceeds the rock stratum water content threshold as the high water content abnormal area, and mark it as an independent layer.

[0010] A further improvement of the technical solution of the present invention lies in: in S3, the process of constructing a geological stress prediction model and outputting the vertical stress distribution matrix of each area includes: Obtain the measured vertical stress in the borehole through a borehole stress gauge , combined with the density distribution heat map in the density distribution characteristics, after normalizing the density value to the interval [0,1], divide the training set and the test set in a ratio of 8:2, and use the support vector regression algorithm to construct a geological stress prediction model; Establish a non-linear mapping relationship between density and vertical stress using a radial basis kernel function. Take the mean square error between the predicted stress and the measured vertical stress in the test set as the objective function, establish a density-stress relationship equation, predict the vertical stress, and optimize the width γ, penalty coefficient C, and loss threshold ϵ in the kernel function using grid search and cross-validation; Input the three-dimensional grid density value of the coal mine geological transparent working face into the geological stress prediction model, predict the vertical stress grid by grid, 3 The low-density area is supplemented by linear interpolation. Set the vertical stress threshold, mark the area where σ is greater than the vertical stress threshold as the high stress abnormal area, generate the vertical stress distribution matrix of each area, and store it as floating-point raster data.

[0011] A further improvement of the technical solution of the present invention lies in: in S4, the process of designing the abnormal area determination condition, outputting the abnormal volume, and performing hierarchical early warning includes: Combine the three-dimensional geometric structure coordinate set, the rock stratum water content matrix of each area, and the vertical stress matrix to design the abnormal area determination condition. The abnormal area determination condition includes defining a composite abnormal area and a single abnormal area. Among them, the composite abnormal area is a grid unit that satisfies the over-standard of the rock stratum water content and the vertical stress and is located in the fracture zone and the broken zone, and the single abnormal area is a grid unit that satisfies one of the over-standard conditions of the rock stratum water content and the vertical stress; Traverse the three-dimensional grid of the transparent coal mine geological working face, screen the grid cells that meet the abnormal area determination conditions, extract the geometric center coordinates of the grid cells corresponding to the composite abnormal area and the single abnormal area, merge the adjacent grid cells with a spatial distance not exceeding 50 cm into a continuous abnormal body, and obtain its abnormal volume V; Classify three-level early warnings according to the abnormal volume. If the composite abnormal area V≥2m 3 , then trigger a first-level early warning and generate a shutdown instruction for the shearer. If the composite abnormal area 0.5≤V<2m 3 , then trigger a second-level early warning, and the shearer reduces its speed to 50% for operation. If the single abnormal area V≥1m 3 , then trigger a third-level audible and visual early warning, and generate an early warning message including the three-dimensional coordinate set of the abnormal area, the abnormal volume, and the early warning level.

[0012] A further improvement of the technical solution of the present invention lies in that: in the S5, the process of synthesizing and displaying a three-dimensional visualization image of the three-dimensional geometric structure, the moisture content distribution of each regional rock formation, and the vertical stress distribution includes: Take the three-dimensional geometric structure coordinate set, the moisture content matrix of each regional rock formation, and the vertical stress matrix as the input physical property parameters. The physical property parameters are spatially aligned based on the three-dimensional grid of the transparent coal mine geological working face, and the index relationship between the geometric vertices and the physical property parameters is established through the coordinate mapping table; Using the ray casting volume rendering technique, emit rays along the viewing angle, penetrate the three-dimensional grid for sampling with a step size of 1 cm, and obtain the moisture content of each regional rock formation and the vertical stress point by point. The color of the moisture content of each regional rock formation is mapped through a blue-red gradient color scale, and a halo special effect is superimposed on the high moisture content abnormal area. The vertical stress is mapped through transparency, and the transparency increases with the increase of stress. The high stress abnormal area is displayed with a semi-transparent red overlay. Integrate the color of the moisture content of each regional rock formation and the transparency of the vertical stress. For each sampling point, introduce the Phong lighting model to calculate the final color, and the color of the abnormal area is preferentially rendered and covered; When new data is triggered, the grid within a radius of 2 m centered on the changed area is resampled and rendered, supporting arbitrary plane slicing to display the moisture content and vertical stress distribution of each regional rock formation, and clicking on the grid cell outputs the three-dimensional coordinates, the moisture content of the rock formation, and the vertical stress value.

[0013] In the second aspect, a construction system for a transparent coal mine geological working face is used to implement a construction method for a transparent coal mine geological working face, including a coal mine geological data acquisition module, a transparent working face construction module, a transparent working face early warning module, and an image synthesis module. Among them, the modules are electrically connected to each other; The coal mine geological data acquisition module collects visible light and near-infrared image data and point cloud coordinate data of the transparent coal mine geological working face, preprocesses them, and extracts the characteristics of the transparent coal mine geological working face; The transparent working face construction module constructs a three-dimensional structure reconstruction model, a rock formation physical property inversion model, and a geological stress prediction model in combination with the characteristics of the transparent working face of coal mine geology, and outputs a three-dimensional geometric structure coordinate set of the transparent working face of coal mine geology, a matrix of water content of rock formations in each area, and a vertical stress distribution matrix; The transparent working face warning module designs abnormal area determination conditions based on the output three-dimensional geometric structure coordinate set, the matrix of water content of rock formations in each area, and the vertical stress distribution matrix, outputs the abnormal volume, and conducts hierarchical warnings; The image synthesis module synthesizes and displays three-dimensional visualization images showing the three-dimensional geometric structure, the water content distribution of rock formations in each area, and the vertical stress distribution.

[0014] Due to the adoption of the above technical solutions, the technical progress achieved by the present invention compared with the prior art is as follows: 1. The present invention provides a construction method and system for a transparent working face of coal mine geology. Through the multi-source data fusion of multi-spectral images and laser point clouds, a three-dimensional geological model with centimeter-level accuracy is constructed, realizing the holographic transparent expression of coal seam structure, water content, and stress distribution, accurately identifying hidden water inrush channels, fracture zones, and stress concentration areas, solving the problems of low spatial resolution and many blind areas in traditional geological exploration, and improving the safety mining efficiency.

[0015] 2. The present invention provides a construction method and system for a transparent working face of coal mine geology. Based on the multi-model collaboration and dynamic data update mechanism, the water content of each rock formation and the vertical stress threshold are monitored in real time, realizing a minute-level warning response for the composite abnormal area, reducing the false alarm rate of accidents such as water inrush and rock burst to less than 5%, and significantly enhancing the timeliness of disaster prevention and control.

[0016] 3. The present invention provides a construction method and system for a transparent working face of coal mine geology. Through the ray casting volume rendering technology, a multi-attribute three-dimensional visualization scene is synthesized, supporting dynamic sectioning, coordinate query, and interactive analysis, providing an intuitive decision-making basis for shearer path planning and roadway support design, reducing the dependence on manual experience, increasing the resource recovery rate by 10%-15%, and promoting the transformation of coal mine mining towards intelligence and data-driven. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention. Example 1, as Figure 1 As shown, the present invention provides a method for constructing a geologically transparent working surface in a coal mine, comprising the following steps: S1. Collect and pre-process the visible light and near-infrared image data and point cloud coordinate data of the transparent working face of the coal mine geology. Arrange explosion-proof multi-spectral imaging equipment at intervals of 20 meters on the roof of the coal mine tunnel. The multi-spectral imaging equipment has a built-in spectroscopic prism component to synchronously collect images of the 400-700nm visible light band and the 700-2500nm near-infrared band. The multi-spectral imaging equipment switches the bands through the filter wheel, dynamically adjusts the exposure time of each band according to the underground light intensity, and outputs visible light and near-infrared image data. Phase-type three-dimensional laser scanning equipment is staggered on both sides of the coal mine tunnel. The three-dimensional laser scanning equipment emits a near-infrared laser with a modulation frequency of 150MHz, obtains the distance to the target point through the phase difference ranging principle, and synchronously records the scanned water The horizontal and elevation angles are used to output point cloud coordinate data. The multispectral imaging device is installed at a height of 2.5 meters from the base plate, and the elevation angle is adjusted to 30°. The scanning head of the 3D laser scanning device is 1.5 meters from the base plate. The horizontal scanning angle covers 120°, and the vertical scanning angle covers -30° to +45°. The 3D laser scanning devices are connected by time synchronization signal lines. The dark current matrix and gain parameters are used to correct the original visible light and near-infrared image data. The image resolution after correction remains at 5472×3648 pixels and is stored in 16-bit RAW format. An adaptive voxel filtering algorithm is used to divide the point cloud space into a 3mm×3mm×3mm voxel grid, and the centroid point with a density of ≥2000 points / square meter in each voxel is retained. S2. Extract the characteristics of the coal mine geological transparent working face from the preprocessed visible and near-infrared image data and point cloud coordinate data. Based on the preprocessed visible light image data, extract the texture uniformity feature. The process of extracting the texture uniformity feature includes using the local binary pattern algorithm to divide the image into local windows of 8×8 pixels, performing gray-scale comparison between the central pixel and neighboring pixels for each window to generate 256-dimensional LBP coding values, and obtaining the variance of the LBP coding values within the window. Mark the continuous region with a variance less than 0.1 as the complete rock stratum boundary coordinate set. Based on the preprocessed near-infrared image data, extract the near-infrared reflectance difference feature. The process of extracting the near-infrared reflectance difference feature includes extracting the reflectance values in the 850nm and 1650nm bands, setting the dual-band reflectance ratio threshold according to the moisture content sensitivity characteristics of underground rocks. If the reflectance ratio between 850nm and 1650nm bands is greater than the dual-band reflectance ratio threshold, it is determined as a high moisture content abnormal area, and output the spatial coordinate matrix of the high moisture content abnormal area. Based on the preprocessed point cloud coordinate data, extract the surface curvature mutation feature and density distribution feature. The process of extracting the surface curvature mutation feature includes searching for 50 neighboring points within a radius of 3cm of each target point, obtaining the local Gaussian curvature based on the difference between the target point and the normal vectors of its neighboring points, and marking the area with a local Gaussian curvature greater than 0.1 mm -1 as the surface curvature mutation area, and output the three-dimensional coordinate sequence of the potential fissure zone. The process of extracting the density distribution feature includes dividing the point cloud space into cubic voxel grids of 10cm³, counting the number of point clouds in each voxel, and taking the ratio of the number of point clouds in each voxel to the volume of the cubic voxel grid as its density. Screen the voxels with a density less than 1500 points / m 3 and determine them as loose and broken zones, and generate the density distribution heat map of the loose and broken zones; S3. Combine the characteristics of the coal mine geological transparent working face to construct a three-dimensional structure reconstruction model, a rock stratum physical property inversion model, and a geological stress prediction model, and output the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face, the matrix of the moisture content of each area of the rock stratum, and the vertical stress distribution matrix. Use the texture uniformity feature, the surface curvature mutation feature, and the point cloud coordinate set with a filtered density ≥ 2000 points per square meter as input data, and adopt the improved iterative closest point algorithm to iteratively solve the optimal rigid body transformation matrix . The iterative solution process includes initializing the rigid body transformation matrix and aligning based on the centroid of the point cloud. Search for the nearest neighbor point of each point in the target point cloud to construct a comprehensive objective function including the basic point cloud registration term , the complete rock stratum boundary constraint term and the fissure zone curvature constraint term . Among them, the basic point cloud registration term minimizes the source point cloud and the target point cloud For the rigid body transformation error, the complete rock formation boundary constraint term forces the complete rock formation boundary points to match the borehole data points, the curvature constraint term of the fracture zone increases the registration weight of the points in the curvature mutation area, key point pairs are screened according to the complete rock formation boundary constraint term and the curvature constraint term of the fracture zone, the optimal rigid body transformation matrix is solved by singular value decomposition, the maximum number of iterations is set, and the iteration stops until the maximum number of iterations is reached, and a three-dimensional structure reconstruction model is generated. The calculation process is as follows: ; ; ; ; ; ; Among them, is the coordinate of the complete rock formation boundary point measured by the borehole, is the texture constraint weight calibrated by 50 registration experiments with known complete rock formation boundaries. The boundary refers to the complete rock formation area screened by the texture uniformity feature, is the curvature constraint weight optimized based on the registration error reduction rate of the fracture zone. The fracture refers to the potential fracture area marked by the curvature mutation feature, is the rotation matrix obtained by SVD, and are the left and right singular vector matrices, both of which are orthogonal matrices, represents the translation vector, and are the weighted centroids of the source point cloud and the target point cloud. The multi-view registration point clouds are fused into a point set in a unified coordinate system. Based on the Delaunay triangulation algorithm, a point cloud surface mesh is constructed, adjacent points are connected to generate triangular patches, and the maximum side length of the triangular patches ≤ 5 cm. After removing the hanging triangular patches, the vertex coordinates of the mesh and the topological relationship of the triangular patches are extracted, and the three-dimensional geometric structure coordinate set of the transparent working face of the coal mine geology is output. The spatial coordinate matrix of the high water content abnormal area in the near-infrared reflectance difference feature is used as the input data. Based on the extension of the Lambert-Beer law, a non-linear relationship between the reflectance ratio R and the rock formation water content w is established, and a rock formation physical property inversion model is generated. The calculation process is as follows: ; Adjust the coefficients of the rock formation physical property inversion model through the Bayesian optimization algorithm, introduce the measured water content samples of the rock formation in boreholes, minimize the root mean square error between the predicted water content value of the rock formation and the measured water content sample value of the rock formation in boreholes, align the spatial coordinate matrix of the high water content abnormal area with the three-dimensional grid of the coal mine geological transparent working face, calculate the reflectivity ratio grid by grid and invert the water content of the rock formation, use Kriging interpolation to fill in the non-abnormal area of the water content of the rock formation, combine the measured water content sample data of the rock formation in boreholes, calibrate the output of the rock formation physical property inversion model through weighted least squares method, generate the water content matrix of each area of the coal mine geological transparent working face, store it as floating-point raster data, set the water content threshold of the rock formation, determine the area where the predicted water content value of the rock formation exceeds the water content threshold of the rock formation as the high water content abnormal area, mark it as an independent layer, and obtain the measured vertical stress of the borehole through the borehole stress gauge , combined with the density distribution heat map in the density distribution characteristics, the density value After normalizing to the interval [0, 1], divide the training set and the test set according to the ratio of 8:2. Adopt the support vector regression algorithm to construct a geological stress prediction model. Use the radial basis kernel function to establish a non-linear mapping relationship between density and vertical stress. Take the mean square error between the predicted stress and the measured vertical stress in the test set as the objective function, establish a density-stress relationship equation, predict the vertical stress, and optimize the width γ, penalty coefficient C, and loss threshold ϵ in the kernel function through grid search and cross-validation. Input the three-dimensional grid density value of the coal mine geological transparent working face into the geological stress prediction model, and predict the vertical stress grid by grid 3 The low-density area is supplemented by linear interpolation. Set the vertical stress threshold, and mark the area where σ is greater than the vertical stress threshold as the high stress abnormal area. Generate the vertical stress distribution matrix of each area and store it as floating-point raster data; S4. Based on the output three-dimensional geometric structure coordinate set, the water content matrix of each area of the rock formation and the vertical stress distribution matrix, design the abnormal area determination conditions, output the abnormal volume, and conduct hierarchical early warning. Combine the three-dimensional geometric structure coordinate set, the water content matrix of each area of the rock formation and the vertical stress matrix to design the abnormal area determination conditions. The abnormal area determination conditions include defining compound abnormal areas and single abnormal areas. Among them, the compound abnormal area is the grid unit that satisfies the over-standard water content and vertical stress of the rock formation and is located in the fracture zone and the broken zone. The single abnormal area is the grid unit that satisfies one of the over-standard conditions of the water content and vertical stress of the rock formation. Traverse the three-dimensional grid of the coal mine geological transparent working face, screen the grid units that meet the abnormal area determination conditions, extract the geometric center coordinates of the grid units corresponding to the compound abnormal area and the single abnormal area, merge the adjacent grid units with a spatial distance not exceeding 50 cm into continuous abnormal bodies, and obtain their abnormal volume V. Divide the three-level early warning according to the abnormal volume. If the compound abnormal area V≥2m 3, then trigger the first level warning, generate the coal mining machine shutdown command, if the composite abnormal area 0.5≤V<2m 3 , then trigger the second level warning, the coal mining machine slows down to 50% operation, if the single abnormal area V ≥ 1m 3 , then a three-level sound and light warning is triggered, generating warning information including the three-dimensional coordinate set of the abnormal area, the abnormal volume and the warning level; S5. Synthesize a 3D visualization image showing the 3D geometric structure, the distribution of rock moisture content in each region, and the vertical stress distribution. The 3D geometric structure coordinate set, the rock moisture content matrix in each region, and the vertical stress matrix are used as input physical parameters. The physical parameters are spatially aligned based on the 3D grid of the coal mine geological transparent working surface. The index relationship between the geometric vertices and the physical parameters is established through the coordinate mapping table. Using the ray casting volume rendering technology, light is emitted along the observation angle, penetrating the 3D grid with a step size of 1 cm to sample the rock moisture content and vertical stress in each region point by point. The color of the rock moisture content in each region is gradiented from blue to red. Color-scale mapping, halo effects superimposed on high water content anomaly areas, vertical stress mapped through transparency, transparency increases with increasing stress, high stress anomaly areas display translucent red superposition, fusion of rock water content color and vertical stress transparency in each area, introduction of Phong lighting model to calculate the final color for each sampling point, priority rendering and coverage of anomaly area color, new data triggers resampling and rendering of grids within a radius of 2m centered on the change area, supports arbitrary plane sectioning to display the distribution of rock water content and vertical stress in each area, click on the grid unit to output the three-dimensional coordinates, rock water content and vertical stress values.

[0020] Example 2, as Figure 1 As shown, based on Example 1, the present invention provides a technical solution: a system for constructing a coal mine geological transparent working surface, used to implement a method for constructing a coal mine geological transparent working surface, including a coal mine geological data acquisition module, a transparent working surface construction module, a clear working surface early warning module and an image synthesis module, wherein the modules are electrically connected; Coal mine geological data acquisition module, which collects visible light and near-infrared image data and point cloud coordinate data of the coal mine geological transparent working surface, pre-processes them, and extracts the characteristics of the coal mine geological transparent working surface; The transparent working face construction module combines the characteristics of the coal mine geological transparent working face to construct a 3D structural reconstruction model, a rock stratum physical property inversion model, and a geological stress prediction model. It outputs the 3D geometric structure coordinate set of the coal mine geological transparent working face, the rock stratum moisture content matrix of each area, and the vertical stress distribution matrix. The transparent working face early warning module designs abnormal area judgment conditions, outputs abnormal volume, and issues graded early warnings based on the output 3D geometric structure coordinate set, the water content matrix of each region's rock formation, and the vertical stress distribution matrix; The image synthesis module synthesizes and displays a three-dimensional visualization image showing the three-dimensional geometric structure, the moisture content distribution of rock formations in each area, and the vertical stress distribution.

[0021] The above is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A construction method for a transparent working face of a coal mine geology, characterized in that The following steps are involved: S1. Collect and pre-process visible light and near-infrared image data and point cloud coordinate data of the transparent working face of the coal mine; S2, extracting the geological transparent working surface features of the coal mine from the pre-processed visible light and near infrared image data and point cloud coordinate data; S3. Based on the characteristics of the coal mine geological transparent working face, a three-dimensional structure reconstruction model, a rock stratum physical property inversion model, and a geological stress prediction model are constructed to output the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face, the rock stratum moisture content matrix of each region, and the vertical stress distribution matrix; S4. Based on the output 3D geometric structure coordinate set, the rock moisture content matrix of each region, and the vertical stress distribution matrix, design abnormal area determination conditions, output abnormal volume, and issue graded warnings; S5. Synthesize a 3D visualization image showing the 3D geometric structure, the water content distribution of the rock layers in each region, and the vertical stress distribution.

2. The construction method of a transparent working face for coal mine geology according to claim 1, characterized in that: In S1, the collection and preprocessing process of visible light and near-infrared image data and point cloud coordinate data of the transparent working surface of the coal mine geology includes: Explosion-proof multispectral imaging equipment is deployed at 20-meter intervals on the roof of the coal mine tunnel. The multispectral imaging equipment has a built-in spectroscopic prism component that simultaneously collects images in the 400-700nm visible light band and the 700-2500nm near-infrared band. The multispectral imaging equipment switches bands using a filter wheel, dynamically adjusts the exposure time of each band based on the underground light intensity, and outputs visible light and near-infrared image data. Phase-shifted 3D laser scanning equipment is staggered on both sides of the coal mine tunnel. The 3D laser scanning equipment emits near-infrared lasers with a modulation frequency of 150MHz, obtains the distance to the target point through the phase difference ranging principle, and simultaneously records the horizontal angle and pitch angle of the scan, and outputs point cloud coordinate data; The raw visible and near-infrared image data were corrected using dark current matrix and gain parameters. The image resolution after correction remained at 5472 × 3648 pixels and was stored in 16-bit RAW format. An adaptive voxel filtering algorithm was used to divide the point cloud space into a 3 mm × 3 mm × 3 mm voxel grid, and the centroid points with a density ≥ 2000 points / m2 within each voxel were retained.

3. The construction method of a transparent working face for coal mine geology according to claim 2, characterized in that: In S2, the process of extracting features from the pre-processed visible light and near-infrared image data and point cloud coordinate data includes: Extracting texture uniformity features based on preprocessed visible light image data. The process of extracting texture uniformity features includes using a local binary pattern algorithm to divide the image into 8×8 pixel local windows, performing a grayscale comparison between the central pixel and the neighboring pixels in each window, generating a 256-dimensional LBP code value, and obtaining the variance of the LBP code value within the window. Continuous regions with a variance less than 0.1 are marked as a complete rock formation boundary coordinate set. Based on the preprocessed near-infrared image data, extract the near-infrared reflectance difference features. The process of extracting the near-infrared reflectance difference features includes extracting the reflectance values in the 850nm and 1650nm bands. According to the sensitivity characteristics of the water content of underground rocks, set the dual-band reflectance ratio threshold. If the reflectance ratio of the 850nm and 1650nm bands is greater than the dual-band reflectance ratio threshold, it is determined as a high water content anomaly area, and the spatial coordinate matrix of the high water content anomaly area is output; Based on the pre-processed point cloud coordinate data, the surface curvature mutation characteristics and density distribution characteristics are extracted; The process of extracting surface curvature mutation features includes searching for 50 neighborhood points within a range of 3 cm radius of each target point, obtaining the local Gaussian curvature based on the difference between the normal vectors of the target point and its neighborhood points, and marking the area where the local Gaussian curvature is greater than 0.1 mm -1 as the surface curvature mutation area, and outputting the three-dimensional coordinate sequence of the potential fracture zone; The process of extracting density distribution features includes dividing the point cloud space into cubic voxel grids with a volume of 10 cm³, counting the number of point clouds in each voxel, and taking the ratio of the number of point clouds in each voxel to the volume of the cubic voxel grid as its density. Voxels with a density less than 1500 points / m 3 are screened out, determined as loose broken zones, and a density distribution heat map of the loose broken zones is generated.

4. A construction method of a transparent working face for coal mine geology according to claim 3, characterized in that: In S3, the process of constructing a three-dimensional structural reconstruction model and outputting a three-dimensional geometric structure coordinate set of the coal mine geological transparent working surface includes: The texture uniformity features, surface curvature mutation features, and the point cloud coordinate set with a density of ≥2000 points / m2 after filtering are used as input data; An improved iterative closest point algorithm is adopted to iteratively solve the optimal rigid body transformation matrix , and the iterative solution process includes initializing the rigid body transformation matrix, aligning based on the centroid of the point cloud, searching for the nearest neighbor points of each point in the target point cloud, and constructing a comprehensive objective function including the basic point cloud registration term , the complete rock stratum boundary constraint term and the fracture zone curvature constraint term . Among them, the basic point cloud registration term minimizes the rigid body transformation error between the source point cloud and the target point cloud . The complete rock stratum boundary constraint term forces the complete rock stratum boundary points to be matched to the borehole data points. The fracture zone curvature constraint term increases the registration weight of the points in the curvature mutation region. Key point pairs are screened according to the complete rock stratum boundary constraint term and the fracture zone curvature constraint term. The optimal rigid body transformation matrix is solved by singular value decomposition. The maximum number of iterations is set, and the iteration stops until the maximum number of iterations is reached, and a three-dimensional structure reconstruction model is generated; Fuse multi-view registered point clouds into a point set in a unified coordinate system, construct a point cloud surface mesh based on the Delaunay triangulation algorithm, connect adjacent points to generate triangular patches, ensure that the maximum side length of the triangular patches is ≤ 5 cm, extract the vertex coordinates of the mesh and the topological relationship of the triangular patches after removing the hanging triangular patches, and output the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face.

5. The construction method of a transparent working face for coal mine geology according to claim 4, characterized in that: In S3, the process of constructing a rock mass physical property inversion model and outputting the water content matrix of each region's rock stratum includes: Use the spatial coordinate matrix of the high water content abnormal area in the near-infrared reflectance difference feature as input data, establish a non-linear relationship between the reflectance ratio R and the water content w of the rock stratum based on the extension of Lambert-Beer's law, and generate a rock mass physical property inversion model; Adjust the coefficients of the rock mass physical property inversion model through the Bayesian optimization algorithm, introduce the measured water content samples of the rock stratum in boreholes, and minimize the root mean square error between the predicted water content value of the rock stratum and the measured water content sample value in boreholes; Align the spatial coordinate matrix of the high water content abnormal area with the three-dimensional grid of the coal mine geological transparent working face, calculate the reflectance ratio grid by grid and invert the water content of the rock stratum, use Kriging interpolation to fill in the non-abnormal areas of the water content of the rock stratum, combine the measured water content sample data of the rock stratum in boreholes, calibrate the output of the rock mass physical property inversion model through weighted least squares method, generate the water content matrix of each region's rock stratum of the coal mine geological transparent working face, store it as floating-point raster data, set the water content threshold of the rock stratum, and determine the area where the predicted water content value of the rock stratum exceeds the water content threshold of the rock stratum as a high water content abnormal area and mark it as an independent layer.

6. The construction method of a transparent working face for coal mine geology according to claim 5, characterized in that: In S3, the process of constructing a geological stress prediction model and outputting the vertical stress distribution matrix of each region includes: Obtain the measured vertical stress in the borehole through a borehole stress gauge , combined with the density distribution heat map in the density distribution characteristics, after normalizing the density values to the interval [0, 1], divide the training set and the test set in an 8:2 ratio, and use the support vector regression algorithm to construct a geological stress prediction model; Establish a non-linear mapping relationship between density and vertical stress using a radial basis kernel function, take the mean square error between the predicted stress and the measured vertical stress in the test set as the objective function, establish a density-stress relationship equation to predict the vertical stress, and optimize the width γ, penalty coefficient C, and loss threshold ϵ in the kernel function using grid search and cross-validation; Input the three-dimensional grid density values of the transparent coal mine working face into the geological stress prediction model, and predict the vertical stress grid by grid. 3 The low-density areas of are supplemented by linear interpolation. Set the vertical stress threshold, mark the areas where σ is greater than the vertical stress threshold as high-stress anomaly areas, generate the vertical stress distribution matrix for each area, and store it as floating-point raster data.

7. A construction method of a transparent working face for coal mine geology according to claim 6, characterized in that: In S4, the process of designing abnormal area determination conditions, outputting the abnormal volume, and performing hierarchical early warning includes: Combine the three-dimensional geometric structure coordinate set, the water content matrix of each region's rock stratum, and the vertical stress matrix to design abnormal area determination conditions. The abnormal area determination conditions include defining a composite abnormal area and a single abnormal area. Among them, the composite abnormal area is a grid unit that satisfies the over-standard of the water content of the rock stratum and the vertical stress and is located in the fissure zone and the broken zone, and the single abnormal area is a grid unit that satisfies one of the over-standard conditions of the water content of the rock stratum and the vertical stress; Traverse the three-dimensional grid of the coal mine geological transparent working face, screen the grid units that meet the abnormal area determination conditions, extract the geometric center coordinates of the grid units corresponding to the composite abnormal area and the single abnormal area, merge adjacent grid units with a spatial distance not exceeding 50 cm into a continuous abnormal body, and obtain its abnormal volume V; Divide into three-level early warnings according to the abnormal volume. If the composite abnormal area V ≥ 2m 3 , trigger a first-level early warning and generate a shutdown instruction for the shearer. If the composite abnormal area 0.5 ≤ V < 2m 3 , trigger a second-level early warning and the shearer runs at a reduced speed of 50%. If the single abnormal area V ≥ 1m 3 , trigger a third-level audible and visual early warning and generate early warning information including the three-dimensional coordinate set of the abnormal area, the abnormal volume, and the early warning level.

8. A construction method of a transparent working face for coal mine geology according to claim 7, characterized in that: In S5, the process of synthesizing and displaying a three-dimensional visualization image of the three-dimensional geometric structure, the water content distribution of each region's rock stratum, and the vertical stress distribution includes: Taking the three-dimensional geometric structure coordinate set, the matrix of water content of rock formations in each area, and the vertical stress matrix as the input physical property parameters, the physical property parameters are spatially aligned based on the three-dimensional grid of the transparent working face of coal mine geology, and the index relationship between geometric vertices and physical property parameters is established through the coordinate mapping table; Using the ray casting volume rendering technology, emit rays along the observation perspective, penetrate the three-dimensional grid for sampling with a step size of 1 cm, and obtain the water content and vertical stress of rock formations in each area point by point. The color of the water content of rock formations in each area is mapped through the blue-red gradient color scale, and a halo special effect is superimposed on the high water content anomaly area. The vertical stress is mapped through transparency, and the transparency increases with the increase of stress. The high stress anomaly area is displayed with a semi-transparent red overlay. Integrate the color of the water content of rock formations in each area and the transparency of the vertical stress. For each sampling point, introduce the Phong lighting model to calculate the final color, and the color of the anomaly area is preferentially rendered and covered; When new data is triggered, re-sample and render the grid within a radius of 2 m centered on the changed area, support arbitrary plane cutting to display the water content and vertical stress distribution of rock formations in each area, and click on the grid cell to output the three-dimensional coordinates, water content of rock formations, and vertical stress values.

9. A construction system for a transparent coal mine geological working face, which is used to implement the construction method of a transparent coal mine geological working face described in any one of claims 1-8 above, characterized in that: It includes a coal mine geological data acquisition module, a transparent working face construction module, a transparent working face warning module, and an image synthesis module. Among them, the modules are electrically connected to each other.

10. The construction system of a transparent working face for coal mine geology according to claim 9, characterized in that: The coal mine geological data acquisition module collects visible light and near-infrared image data and point cloud coordinate data of the transparent working face of coal mine geology, preprocesses them, and extracts the characteristics of the transparent working face of coal mine geology; The transparent working face construction module constructs a three-dimensional structure reconstruction model, a rock formation physical property inversion model, and a geological stress prediction model in combination with the characteristics of the transparent working face of coal mine geology, and outputs the three-dimensional geometric structure coordinate set, the matrix of water content of rock formations in each area, and the vertical stress distribution matrix of the transparent working face of coal mine geology; The transparent working face warning module designs the determination conditions for abnormal areas based on the output three-dimensional geometric structure coordinate set, the matrix of water content of rock formations in each area, and the vertical stress distribution matrix, outputs the abnormal volume, and conducts hierarchical warnings; The image synthesis module synthesizes and displays a three-dimensional visualization image of the three-dimensional geometric structure, the water content distribution of rock formations in each area, and the vertical stress distribution.

Images