A method and system for constructing a transparent working face of a coal mine geological section

Abstract

The application discloses a kind of construction method and system of coal mine geological transparent working face, it is related to intelligent mine and geological information technical field, including the following steps: acquisition pre-processing visible light, near-infrared image and point cloud data, extract texture, reflectivity, curvature, density feature, construct three-dimensional structure reconstruction model, rock layer physical property inversion model and geological stress prediction model, generate three-dimensional geometric structure coordinate set, each area rock layer moisture content matrix and vertical stress distribution matrix, set abnormal determination condition, output abnormal area three-dimensional coordinates and hierarchical early warning, synthesis three-dimensional visualization image, realize working face geological state holographic transparent expression.The application realizes that coal mine geological transparent working face centimeter level three-dimensional reconstruction, concealed water-bearing area and stress abnormal area accurate positioning, constructs full attribute dynamic visualization scene, significantly improves water inrush, rock burst and other disaster early warning timeliness and accuracy, provides real-time geological decision support for intelligent mining, guarantees safe and efficient production.

Description

Technical Field

[0001] The present invention relates to the field of intelligent mines and geological information technology, and in particular to a method and system for constructing a geologically transparent working face in a coal mine. Background Art

[0002] Coal geology is a geological discipline that studies the formation, distribution and mining conditions of coal deposits. It mainly involves the occurrence patterns of coal seams, surrounding rock characteristics, structural features and hydrogeological conditions. It evaluates coal resource reserves, mining feasibility and potential disaster risks by analyzing the stratigraphic structure, physical and mechanical properties of coal rocks, the degree of development of faults and folds, and the patterns of groundwater migration. During the coal mine exploration stage, coal mine geology determines the thickness, inclination and stability of coal seams. During the mining process, it monitors the stability of the roof and floor, gas occurrence and water inrush channels to guide the layout and support design of tunnels. At the same time, coal mine geology combines geophysical detection, drilling data and three-dimensional modeling technology to construct a geological model of the mining area, providing a scientific basis for gas extraction, water hazard prevention and control, and rock burst warning. As a bridge connecting resource development and safe production, coal mine geology supports intelligent mining systems to achieve precise geological navigation and risk prevention by revealing the interaction patterns between geological bodies and mining activities. It is a basic supporting discipline for the sustainable development of the modern coal industry.

[0003] To address the lack of transparency in geological information and delayed disaster warnings in coal mining, existing technologies use a combination of single-hole sampling and static geological modeling. However, this approach suffers from low data coverage density and difficulty fusing heterogeneous multi-source data. This leads to a high rate of missed detection of hidden disaster-causing objects and poor timeliness in model updates. To address these limitations, a method and system for constructing a geologically transparent working surface in coal mines was proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for constructing a geologically transparent working face in a coal mine, so as to solve the problems raised in the above-mentioned background technology.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is as follows: In a first aspect, a method for constructing a geologically transparent working face in a coal mine comprises the following steps:

[0006] 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;

[0007] 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;

[0008] 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;

[0009] 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;

[0010] 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.

[0011] 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:

[0012] 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.

[0013] 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;

[0014] 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.

[0015] The raw visible and near-infrared image data were corrected using the dark current matrix and gain parameters. The image resolution after correction remained at 5472 × 3648 pixels and was stored in 16-bit RAW format.

[0016] 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.

[0017] 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:

[0018] 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.

[0019] Based on the pre-processed near-infrared image data, the near-infrared reflectivity difference feature is extracted. The process of extracting the near-infrared reflectivity difference feature includes extracting the reflectivity values ​​of the 850nm and 1650nm bands, setting the dual-band reflectivity ratio threshold according to the sensitive characteristics of the water content of the downhole rock, and if the reflectivity ratio of the 850nm and 1650nm bands is greater than If the value is greater than the dual-band reflectivity ratio threshold, it is determined to be a high water content abnormal area, and the spatial coordinate matrix of the high water content abnormal area is output;

[0020] Based on the pre-processed point cloud coordinate data, the surface curvature mutation characteristics and density distribution characteristics are extracted;

[0021] The process of extracting surface curvature mutation features includes searching for 50 neighboring points within a radius of 3 cm for each target point, obtaining the local Gaussian curvature based on the difference in the normal vectors of the target point and its neighboring points, and marking the local Gaussian curvature greater than 0.1 mm. -1 The area is the surface curvature mutation area, and the three-dimensional coordinate sequence of the potential fracture zone is output;

[0022] The process of extracting density distribution features includes dividing the point cloud space into 10 cm³ cubic voxel grids, 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. The density is screened out to be less than 1500 points / m 3 The voxels are determined to be loose fracture zones, and a density distribution heat map of the loose fracture zones is generated.

[0023] 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 a three-dimensional geometric structure coordinate set of the coal mine geological transparent working surface includes:

[0024] 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;

[0025] Use the improved iterative closest point algorithm to iteratively solve the optimal rigid body transformation matrix , the iterative solving process comprises initializing a rigid body transformation matrix, searching for a nearest neighbor of each point in the target point cloud based on point cloud centroid alignment, constructing a comprehensive objective function containing a basic point cloud registration term , a complete rock boundary constraint term and a fracture zone curvature constraint term , wherein the basic point cloud registration term minimizes the rigid body transformation error of the source point cloud and the target point cloud , the complete rock boundary constraint term forces the complete rock boundary points to match the borehole data points, and the fracture zone curvature constraint term enhances the registration weight of points in the curvature mutation area, the key point pairs are screened according to the complete rock 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 is stopped when the maximum number of iterations is reached, and a three-dimensional structure reconstruction model is generated, and the calculation process is as follows:

[0026] ;

[0027] ;

[0028] ;

[0029] ;

[0030] ;

[0031] ;

[0032] wherein, is the coordinate of the complete rock boundary point measured by the borehole, is the texture constraint weight calibrated by 50 groups of known complete rock boundary registration experiments, the boundary refers to the complete rock area selected by the texture uniformity feature, is the curvature constraint weight determined based on the fracture zone registration error reduction rate optimization, the fracture refers to the potential fracture area marked by the curvature mutation feature, is the rotation matrix obtained by SVD, and are left and right singular vector matrices, both of which are orthogonal matrices, represents a translation vector, and are the weighted centroids of the source point cloud and the target point cloud;

[0033] The multi-view registration point cloud is fused into a point set in a unified coordinate system, a point cloud surface grid is constructed based on a Delaunay triangulation algorithm, triangular facets are generated by connecting adjacent points, the maximum edge length of the triangular facets is less than or equal to 5 cm, after the suspended triangular facets are removed, the grid vertex coordinates and the triangular facet topological relationship are extracted, and a three-dimensional geometric structure coordinate set of a coal mine geological transparent working face is output.

[0034] Further improvement of the technical scheme of the present application is that in the S3, the process of constructing the rock mass property inversion model and outputting the rock mass water content matrix of each region comprises:

[0035] The spatial coordinate matrix of the high water content abnormal area in the near-infrared reflectivity difference feature is taken as input data, a nonlinear relationship between the reflectivity ratio R and the rock mass water content w is established based on the extension of the Lambert-Beer law, and a rock mass property inversion model is generated, and the calculation process is as follows:

[0036] ;

[0037] The rock mass property inversion model coefficient is adjusted through the Bayesian optimization algorithm, the measured rock mass water content sample of the borehole is introduced, and the root mean square error between the rock mass water content prediction value and the measured rock mass water content sample value is minimized;

[0038] The spatial coordinate matrix of the high water content abnormal area is aligned with the three-dimensional grid of the coal mine geological transparent working face, the reflectivity ratio is calculated grid by grid, and the rock mass water content is inverted, the rock mass water content non-abnormal area is filled by using the Kriging interpolation, the rock mass property inversion model output is calibrated by using the weighted least square method in combination with the measured rock mass water content sample data, the rock mass water content matrix of each region of the coal mine geological transparent working face is generated, and is stored as a floating point grid data, the rock mass water content threshold is set, the area with a rock mass water content prediction value exceeding the rock mass water content threshold is determined as a high water content abnormal area, and is marked as an independent layer.

[0039] Further improvement of the technical scheme of the present application is that in the S3, the process of constructing the rock mass property inversion model and outputting the rock mass water content matrix of each region comprises:

[0040] The vertical stress of the borehole is obtained by using a borehole stress meter , the density value is normalized to the interval [0, 1] after the density distribution feature is combined with the density distribution thermogram, the training set and the test set are divided in a ratio of 8:2, the support limited regression algorithm is used, and the geological stress prediction model is constructed;

[0041] The radial basis kernel function is used to establish a nonlinear mapping relationship between density and vertical stress. The mean square error between the predicted stress in the test set and the measured vertical stress is used as the objective function to establish a density-stress relationship equation and predict the vertical stress. Grid search and cross-validation are used to optimize the width γ, penalty coefficient C, and loss threshold ϵ in the kernel function.

[0042] The three-dimensional grid density value of the coal mine geological transparent working face is input into the geological stress prediction model, and the vertical stress is predicted grid by grid. 3 The low-density areas are supplemented by linear interpolation, and the vertical stress threshold is set. The areas where σ is greater than the vertical stress threshold are marked as high stress anomaly areas. The vertical stress distribution matrix of each area is generated and stored as floating-point raster data.

[0043] A further improvement of the technical solution of the present invention is that in S4, the process of designing abnormal area determination conditions, outputting abnormal volume, and performing graded warning includes:

[0044] Combining the three-dimensional geometric structure coordinate set, the rock moisture matrix and the vertical stress matrix of each region, the abnormal area determination conditions are designed. The abnormal area determination conditions include defining composite abnormal areas and single abnormal areas. Composite abnormal areas are grid cells that meet the conditions of rock moisture content and vertical stress exceeding the standard and are located in fracture zones and fracture zones. Single abnormal areas are grid cells that meet the conditions of rock moisture content and vertical stress exceeding the standard.

[0045] Traverse the three-dimensional grid of the coal mine geological transparent working face, screen the grid cells that meet the abnormal area judgment 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 of no more than 50 cm into a continuous abnormal body, and obtain its abnormal volume V;

[0046] The three-level warning is divided according to the abnormal volume. If the composite 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.

[0047] A further improvement of the technical solution of the present invention is that in S5, the process of synthesizing a three-dimensional visualization image showing the three-dimensional geometric structure, the water content distribution of each region of the rock layer, and the vertical stress distribution includes:

[0048] The three-dimensional geometric structure coordinate set, the water content matrix of each regional rock layer, and the vertical stress matrix are used as input physical property parameters. The physical property parameters are spatially aligned based on the three-dimensional grid of the coal mine geological transparent working face, and the index relationship between the geometric vertices and the physical property parameters is established through the coordinate mapping table.

[0049] Using raycasting volume rendering technology, light is emitted along the observation angle, penetrating the three-dimensional grid sampling with a step size of 1 cm, and the water content and vertical stress of each area of ​​the rock formation are obtained point by point. The color of the rock formation water content in each area is mapped using a blue-red gradient color scale. High water content abnormal areas are superimposed with a halo effect. Vertical stress is mapped using transparency. The transparency increases with increasing stress. High stress abnormal areas are displayed as a semi-transparent red overlay. The rock formation water content color of each area and the vertical stress transparency are integrated. For each sampling point, the Phong lighting model is introduced to calculate the final color, and the color of the abnormal area is rendered and overlaid first.

[0050] New data triggers the resampling and rendering of the grid within a 2m radius centered on the changed area. It supports arbitrary plane sectioning to display the rock moisture content and vertical stress distribution in each area. Click on the grid cell to output the three-dimensional coordinates, rock moisture content and vertical stress values.

[0051] In a second aspect, a system for constructing a geologically transparent working face in a coal mine is provided, which is used to implement a method for constructing a geologically transparent working face in a coal mine, and includes a coal mine geological data acquisition module, a transparent working face construction module, a clear working face warning module, and an image synthesis module, wherein the modules are electrically connected to each other;

[0052] The coal mine geological data acquisition module 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;

[0053] The transparent working face construction module, in combination with the characteristics of the coal mine geological transparent working face, constructs a three-dimensional structure reconstruction model, a rock stratum physical property inversion model and a geological stress prediction model, and outputs a three-dimensional geometric structure coordinate set of the coal mine geological transparent working face, a rock stratum moisture content matrix in each region and a vertical stress distribution matrix;

[0054] The transparent working face warning module designs abnormal area determination conditions based on the output three-dimensional geometric structure coordinate set, the rock formation water content matrix of each area, and the vertical stress distribution matrix, outputs the abnormal volume, and issues graded warnings;

[0055] The image synthesis module synthesizes a three-dimensional visual image that displays the three-dimensional geometric structure, the water content distribution of the rock layers in each region, and the vertical stress distribution.

[0056] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0057] 1. The present invention provides a method and system for constructing a geologically transparent working face in a coal mine. By fusing multi-source data of multispectral images and laser point clouds, a three-dimensional geological model with centimeter-level accuracy is constructed, achieving a holographic and transparent expression of coal seam structure, moisture content, and stress distribution, accurately identifying hidden water inrush channels, fracture zones, and stress concentration areas, solving the problems of low spatial resolution and multiple blind spots in traditional geological exploration, and improving safe mining efficiency.

[0058] 2. The present invention provides a method and system for constructing a geologically transparent working face in a coal mine. Based on a multi-model collaboration and dynamic data update mechanism, the system monitors the moisture content and vertical stress threshold of each rock layer in real time, achieves minute-level early warning response in complex abnormal areas, reduces the false alarm rate of accidents such as water inrush and rock burst to below 5%, and significantly enhances the timeliness of disaster prevention and control.

[0059] 3. The present invention provides a method and system for constructing a transparent geological working face in a coal mine. It synthesizes a multi-attribute three-dimensional visualization scene through ray casting volume rendering technology, supports dynamic sectioning, coordinate query and interactive analysis, provides an intuitive decision-making basis for coal mining machine path planning and tunnel support design, reduces reliance on manual experience, improves resource recovery rate by 10%-15%, and promotes the transformation of coal mining to intelligent and data-driven. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0061] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0062] 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.

[0063] 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:

[0064] 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.

[0065] S2. Extract the geological transparent working face features of the coal mine from the pre-processed visible light and near-infrared image data and point cloud coordinate data. Extract the texture uniformity features based on the pre-processed visible light image data. The process of extracting the texture uniformity features includes using a local binary pattern algorithm to divide the image into a local window of 8×8 pixels, performing grayscale comparison between the central pixel and the neighboring pixel of each window, generating a 256-dimensional LBP coding value, and obtaining the variance of the LBP coding value in the window. The continuous area with a variance less than 0.1 is marked as a complete rock formation boundary coordinate set. Based on the pre-processed near-infrared image data, extract the near-infrared reflectivity difference features. The process of extracting the near-infrared reflectivity difference features includes extracting the reflectivity values ​​of the 850nm and 1650nm bands, setting a dual-band reflectivity ratio threshold according to the sensitive characteristics of the water content of the underground rock, and if the reflectivity ratio of the 850nm and 1650nm bands is greater than 0.1, the reflectivity ratio of the 850nm and 1650nm bands is greater than 0.1. If the reflectivity is greater than the dual-band reflectivity ratio threshold, it is determined to be a high water content abnormal area, and the spatial coordinate matrix of the high water content abnormal area is output. Based on the preprocessed point cloud coordinate data, the surface curvature mutation characteristics and density distribution characteristics are extracted. The process of extracting the surface curvature mutation characteristics includes searching for 50 neighboring points within a radius of 3 cm for each target point, obtaining the local Gaussian curvature based on the difference in the normal vector of the target point and its neighboring points, and marking the local Gaussian curvature greater than 0.1 mm. -1The area with a sudden change in surface curvature is the output of the three-dimensional coordinate sequence of the potential fracture zone. The process of extracting density distribution features includes dividing the point cloud space into 10 cm³ cubic voxel grids, 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. The density is screened to be less than 1500 points / m 3 The voxels are determined to be loose fracture zones, and a density distribution heat map of the loose fracture zones is generated;

[0066] S3. Combined with 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. 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 are output. The texture uniformity characteristics, surface curvature mutation characteristics, and the point cloud coordinate set with a density of ≥2000 points / square meter after filtering are used as input data. The improved iterative closest point algorithm is used to iteratively solve the optimal rigid body transformation matrix. The iterative solution process includes initializing the rigid body transformation matrix, and based on the point cloud centroid alignment, searching for the nearest neighbor of each point in the target point cloud, and constructing a basic point cloud registration term. , complete rock formation boundary constraint and fracture zone curvature constraint The comprehensive objective function of the point cloud registration term is minimized by the source point cloud With the target point cloud The rigid body transformation error is calculated. The complete rock formation boundary constraint item forces the complete rock formation boundary points to match the borehole data points. The fracture zone curvature constraint item increases the registration weight of the points in the curvature mutation area. Key point pairs are selected based on the complete rock formation boundary constraint item and the fracture zone curvature constraint item. The optimal rigid body transformation matrix is ​​solved by singular value decomposition. The maximum number of iterations is set. The iteration is stopped when the maximum number of iterations is reached to generate a 3D structural reconstruction model. The calculation process is as follows:

[0067] ;

[0068] ;

[0069] ;

[0070] ;

[0071] ;

[0072] ;

[0073] in, is the coordinate of the complete rock formation boundary point measured by drilling, To calibrate the texture constraint weight through 50 sets of known complete rock boundary registration experiments, the boundary refers to the complete rock area selected by the texture uniformity feature, To optimize the curvature constraint weight determined based on the fracture zone registration error reduction rate, the fracture refers to the potential fracture area marked by the curvature mutation feature, is the rotation matrix obtained by SVD, and are 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, and the multi-view registration point cloud is fused into a point set in a unified coordinate system. Based on the Delaunay triangulation algorithm, a point cloud surface grid is constructed, adjacent points are connected to generate triangular facets, and the maximum edge length of the triangular facets is ≤5cm. After removing the suspended triangular facets, the grid vertex coordinates and triangular facet topological relations are extracted, and the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face is output. The spatial coordinate matrix of the high water content abnormal area in the near-infrared reflectivity difference feature is taken as input data, and based on the extension of the Lambert-Beer law, a nonlinear relationship between the reflectivity ratio R and the rock water content w is established to generate a rock property inversion model, and the calculation process is as follows:

[0074] ;

[0075] The rock property inversion model coefficient is adjusted by the Bayesian optimization algorithm, the measured rock water content sample in the borehole is introduced, and the root mean square error between the rock water content prediction value and the measured rock water content sample value is minimized. The spatial coordinate matrix of the high water content abnormal area is aligned with the three-dimensional grid of the coal mine geological transparent working face, the reflectivity ratio is calculated grid by grid, and the rock water content is inverted. The rock water content non-abnormal area is filled by Kriging interpolation, combined with the measured rock water content sample data in the borehole, the output of the rock property inversion model is calibrated by the weighted least squares method, and the rock water content matrix of each area of the coal mine geological transparent working face is generated. Store as a floating-point grid data, set the rock water content threshold, and determine the area where the rock water content prediction value exceeds the rock water content threshold as the high water content abnormal area, and mark it as an independent layer. The vertical stress measured by the borehole stress meter is obtained, combined with the density distribution thermodynamic map in the density distribution feature, and the density value After normalization to the interval [0,1], the training set and the test set were divided into a ratio of 8:2. The support limited regression algorithm was used to construct a geological stress prediction model. The radial basis kernel function was used to establish a nonlinear mapping relationship between density and vertical stress. The mean square error between the predicted stress and the measured vertical stress in the test set was used as the objective function. The density-stress relationship equation was established to predict the vertical stress. The width γ, penalty coefficient C, and loss threshold ϵ in the kernel function were optimized using grid search and cross-validation. The three-dimensional grid density value of the transparent working face of the coal mine was input into the geological stress prediction model, and the vertical stress was predicted grid by grid. 3 The low-density areas are supplemented by linear interpolation, and the vertical stress threshold is set. The areas where σ is greater than the vertical stress threshold are marked as high stress anomaly areas. The vertical stress distribution matrix of each area is generated and stored as floating-point raster data.

[0076] S4. Based on the output three-dimensional geometric structure coordinate set, the rock moisture matrix of each region and the vertical stress distribution matrix, design the abnormal area judgment conditions, output the abnormal volume, and perform graded warning. Combined with the three-dimensional geometric structure coordinate set, the rock moisture matrix of each region and the vertical stress matrix, design the abnormal area judgment conditions. The abnormal area judgment conditions include defining composite abnormal areas and single abnormal areas. Among them, the composite abnormal area is the grid unit that meets the rock moisture content and vertical stress exceeding the standard and is located in the fracture zone and the fracture zone. The single abnormal area is the grid unit that meets the rock moisture content and vertical stress exceeding the standard. Traverse the three-dimensional grid of the coal mine geological transparent working face, screen the grid units that meet the abnormal area judgment conditions, extract the geometric center coordinates of the grid units corresponding to the composite abnormal area and the single abnormal area, merge the adjacent grid units with a spatial distance of no more than 50cm into a continuous abnormal body, and obtain its abnormal volume V. Divide the warning into three levels according to the abnormal volume. If the composite 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;

[0077] 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.

[0078] 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;

[0079] 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;

[0080] 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.

[0081] 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;

[0082] Image synthesis module, which synthesizes three-dimensional visualization images showing three-dimensional geometric structure, water content distribution of rock layers in various regions, and vertical stress distribution.

[0083] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for constructing a geologically transparent working surface in a coal mine, 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 3D structural reconstruction model, a rock stratum physical property inversion model, and a geological stress prediction model are constructed. The 3D 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 are output, including: The spatial coordinate matrix of the high water content abnormal area in the near-infrared reflectivity difference characteristics is used as input data. Based on the expansion of the Lambert-Beer law, a nonlinear relationship between the reflectivity ratio R and the water content w of the rock formation is established to generate a rock property inversion model. The Bayesian optimization algorithm is used to adjust the coefficients of the rock formation physical property inversion model and introduce the measured rock formation water content samples in the borehole to minimize the root mean square error between the predicted rock formation water content and the measured rock formation water content sample values ​​in the borehole. The spatial coordinate matrix of the high-water-content abnormal area is aligned with the three-dimensional grid of the coal mine geological transparent working face. The reflectivity ratio is calculated grid by grid and the rock moisture content is inverted. Kriging interpolation is used to fill in the non-abnormal rock moisture content area. Combined with the measured rock moisture content sample data of the borehole, the output of the rock physical property inversion model is calibrated using the weighted least squares method to generate the rock moisture content matrix of each area of ​​the coal mine geological transparent working face. The matrix is ​​stored as floating-point raster data, and a rock moisture content threshold is set. The area where the rock moisture content prediction value exceeds the rock moisture content threshold is determined as a high-water-content abnormal area and marked as an independent layer. 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.

2. The method for constructing a transparent working surface in a coal mine 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 the 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 method for constructing a transparent working surface in a coal mine 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 pre-processed near-infrared image data, the near-infrared reflectivity difference feature is extracted. The process of extracting the near-infrared reflectivity difference feature includes extracting the reflectivity values ​​of the 850nm and 1650nm bands, setting the dual-band reflectivity ratio threshold according to the sensitive characteristics of the water content of the downhole rock, and if the reflectivity ratio of the 850nm and 1650nm bands is greater than If the value is greater than the dual-band reflectivity ratio threshold, it is determined to be a high water content abnormal area, and the spatial coordinate matrix of the high water content abnormal 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 neighboring points within a radius of 3 cm for each target point, obtaining the local Gaussian curvature based on the difference in the normal vectors of the target point and its neighboring points, and marking the local Gaussian curvature greater than 0.1 mm. -1 The area is the surface curvature mutation area, and the three-dimensional coordinate sequence of the potential fracture zone is output; The process of extracting density distribution features involves dividing the point cloud space into 10cm 3 The number of point clouds in each voxel is counted, and the ratio of the number of point clouds in each voxel to the volume of the cubic voxel grid is used as its density. The density is less than 1500 points / m 3 The voxels are determined to be loose fracture zones, and a density distribution heat map of the loose fracture zones is generated.

4. The method for constructing a transparent working surface in a coal mine 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; Use 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 based on the point cloud centroid alignment, searching for the nearest neighbor of each point in the target point cloud, and constructing a basic point cloud registration term. , complete rock formation boundary constraint and fracture zone curvature constraint The comprehensive objective function of the point cloud registration term is minimized by the source point cloud With the target point cloud The rigid body transformation error is calculated. The complete rock formation boundary constraint item forces the complete rock formation boundary points to match the drill hole data points. The fracture zone curvature constraint item increases the registration weight of the points in the curvature mutation area. Key point pairs are selected based on the complete rock formation boundary constraint item and the fracture zone curvature constraint item. The optimal rigid body transformation matrix is ​​solved by singular value decomposition. The maximum number of iterations is set. The iteration stops when the maximum number of iterations is reached to generate a 3D structural reconstruction model. The multi-view registration point clouds are fused into a point set in a unified coordinate system. The point cloud surface mesh is constructed based on the Delaunay triangulation algorithm. Adjacent points are connected to generate triangular facets with a maximum side length of ≤5cm. After removing the dangling triangular facets, the topological relationship between the mesh vertex coordinates and the triangular facets is extracted to output the three-dimensional geometric structure coordinate set of the coal mine geological transparent working face.

5. The method for constructing a transparent working surface in a coal mine according to claim 4, 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 vertical stress measured in the borehole through the borehole stress gauge , combined with the density distribution heat map in the density distribution feature, the density value After normalization to the interval [0,1], the training set and the test set were divided into 8:2 ratios, and the geological stress prediction model was constructed using the support-limited regression algorithm; The radial basis kernel function is used to establish a nonlinear mapping relationship between density and vertical stress. The mean square error between the predicted stress in the test set and the measured vertical stress is used as the objective function to establish a density-stress relationship equation and predict the vertical stress. Grid search and cross-validation are used to optimize the width γ, penalty coefficient C, and loss threshold ϵ in the kernel function. The three-dimensional grid density value of the coal mine geological transparent working face is input into the geological stress prediction model, and the vertical stress is predicted grid by grid. 3 The low-density areas are supplemented by linear interpolation, and the vertical stress threshold is set. The areas where σ is greater than the vertical stress threshold are marked as high stress anomaly areas. The vertical stress distribution matrix of each area is generated and stored as floating-point raster data.

6. The method for constructing a transparent working surface in a coal mine according to claim 5, characterized in that: In S4, the process of designing abnormal area determination conditions, outputting abnormal volume, and performing graded warning includes: Combining the three-dimensional geometric structure coordinate set, the rock moisture matrix and the vertical stress matrix of each region, the abnormal area determination conditions are designed. The abnormal area determination conditions include defining composite abnormal areas and single abnormal areas. Composite abnormal areas are grid cells that meet the conditions of rock moisture content and vertical stress exceeding the standard and are located in fracture zones and fracture zones. Single abnormal areas are grid cells that meet the conditions of rock moisture content and vertical stress exceeding the standard. Traverse the three-dimensional grid of the coal mine geological transparent working face, screen the grid cells that meet the abnormal area judgment 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 of no more than 50 cm into a continuous abnormal body, and obtain its abnormal volume V; The three-level warning is divided according to the abnormal volume. If the composite 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.

7. The method for constructing a transparent working surface in a coal mine according to claim 6, characterized in that: In S5, the process of synthesizing a three-dimensional visualization image showing the three-dimensional geometric structure, the water content distribution of each region of the rock layer, and the vertical stress distribution includes: The three-dimensional geometric structure coordinate set, the water content matrix of each regional rock layer, and the vertical stress matrix are used as input physical property parameters. The physical property parameters are spatially aligned based on the three-dimensional grid of the coal mine geological transparent working face, and the index relationship between the geometric vertices and the physical property parameters is established through the coordinate mapping table. Using raycasting volume rendering technology, light is emitted along the observation angle, penetrating the three-dimensional grid sampling with a step size of 1 cm, and the water content and vertical stress of each area of ​​the rock formation are obtained point by point. The color of the rock formation water content in each area is mapped using a blue-red gradient color scale. High water content abnormal areas are superimposed with a halo effect. Vertical stress is mapped using transparency. The transparency increases with increasing stress. High stress abnormal areas are displayed as a semi-transparent red overlay. The rock formation water content color of each area and the vertical stress transparency are integrated. For each sampling point, the Phong lighting model is introduced to calculate the final color, and the color of the abnormal area is rendered and overlaid first. New data triggers the resampling and rendering of the grid within a 2m radius centered on the changed area. It supports arbitrary plane sectioning to display the rock moisture content and vertical stress distribution in each area. Click on the grid cell to output the three-dimensional coordinates, rock moisture content and vertical stress values.

8. A system for constructing a geologically transparent working surface in a coal mine, for implementing the method for constructing a geologically transparent working surface in a coal mine according to any one of claims 1 to 7, characterized in that: It includes a coal mine geological data acquisition module, a transparent working face construction module, a clear working face early warning module and an image synthesis module, wherein the modules are connected with electrical signals.

9. The system for constructing a coal mine geological transparent working surface according to claim 8, 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 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, in combination with the characteristics of the coal mine geological transparent working face, constructs a three-dimensional structure reconstruction model, a rock stratum physical property inversion model and a geological stress prediction model, and outputs a three-dimensional geometric structure coordinate set of the coal mine geological transparent working face, a rock stratum moisture content matrix in each region 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 rock formation water content matrix of each area, and the vertical stress distribution matrix, outputs the abnormal volume, and issues graded warnings; The image synthesis module synthesizes a three-dimensional visual image that displays the three-dimensional geometric structure, the water content distribution of the rock layers in each region, and the vertical stress distribution.

Images