A deposit spatial structure reconstruction method and system based on three-dimensional modeling
By employing a progressive processing mechanism based on the strike and dip of the ore body, the problem of ore body morphology deviation in existing technologies has been solved, enabling precise reconstruction of the spatial structure of the ore deposit and improving the accuracy and reliability of ore body morphology prediction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-03-05
- Publication Date
- 2026-06-09
Smart Images
Figure CN122176213A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of three-dimensional reconstruction technology, specifically relating to a method and system for reconstructing the spatial structure of mineral deposits based on three-dimensional modeling. Background Technology
[0002] Reconstructing the three-dimensional spatial structure of mineral deposits is a core step in geological and mineral exploration and development that determines the accuracy of resource evaluation and the reliability of mining design. It directly relates to the accuracy of ore body boundary delineation and the reliability of reserve calculation. In actual mining areas, geological bodies often exhibit complex characteristics of curvature, folds, faults, and pinch-out recurrence, urgently requiring the use of three-dimensional methods to transform dispersed borehole and profile information into a continuous spatial form.
[0003] Most existing methods use fixed grids or global interpolation for modeling. While this approach of processing all data at once is simple to operate, it is difficult to adapt to the drastic changes in the strike and dip directions of ore bodies. Geological interfaces often exhibit sharp bends or abrupt changes in thickness in space. If a unified computational framework is used, local anomalies will be smoothed out, leading to incorrect connections or separations of the same strata in adjacent sections, and consequently causing systematic deviations in the morphology of the entire ore body.
[0004] The root cause of this bias lies in the obvious spatial continuity and directional dependence of geological strata. Strata maintain relatively stable occurrence and thickness relationships along the direction of ore body extension, while existing methods lack a reconstruction mechanism that progresses step by step along a specific direction. When encountering strata intersections, fault displacements, or abrupt lithological changes, the algorithm cannot automatically determine how subsequent interfaces should continue based on the confirmed parts. This often leads to the same geological unit being repeatedly identified or missed in different sections, especially in blind areas with large borehole spacing and sparse data. Such directional tracking interruptions can directly cause unreasonable spatial breaks or overlaps in the ore body.
[0005] Therefore, how to introduce a progressive processing mechanism along the strike or dip of the ore body during the three-dimensional reconstruction process, so that the geological strata can be continuously and accurately extended to the unexplored areas based on the established parts, without being disturbed by the smoothing effect brought about by global calculation, has become the key issue for achieving accurate reconstruction of the spatial structure of complex ore deposits. Summary of the Invention
[0006] To address the problems existing in the prior art, this invention provides a method and system for reconstructing the spatial structure of mineral deposits based on three-dimensional modeling. The aim is to introduce a progressive processing mechanism along the strike or dip of the ore body, so that geological strata can be continuously and accurately extended to unexplored areas based on the established parts, without being disturbed by the smoothing effect brought about by global calculation, thereby achieving accurate reconstruction of the spatial structure of complex mineral deposits.
[0007] To achieve the above objectives, the present invention provides the following solution: A method for reconstructing the spatial structure of a mineral deposit based on 3D modeling, the method comprising: Collect strike and dip information of ore bodies within the ore deposit, and based on the strike and dip information of ore bodies, divide the ore deposit space into a continuous sequence to obtain the initial stratigraphic sequence; Local geological features are extracted based on the initial stratigraphic sequence, and thickness variation trend data are obtained based on the local geological features. Based on thickness variation trend data, a local three-dimensional mesh is generated to obtain the preliminary ore body interface; Based on the preliminary ore body interface tracing direction extension, the predicted point set of the unexplored area is obtained to obtain the extension path; Based on the extended path, the volume element filling method is used to connect the disconnected areas and obtain the complete ore body structure; Based on the verification of the continuity of the dip direction of the complete ore body structure, the overlapping areas are identified, and the final three-dimensional reconstruction model is obtained.
[0008] Preferably, the method for collecting strike and dip information of ore bodies within the ore deposit, and then dividing the ore deposit space into a continuous sequence based on the strike and dip information of the ore bodies to obtain the initial stratigraphic sequence includes: By collecting borehole data and profile data of the ore deposit, the strike and dip information of the ore body within the ore deposit can be obtained; Based on the strike and dip information of the ore body, a preliminary spatial distribution record of the ore deposit is obtained; Based on the preliminary spatial distribution record, the space of the ore deposit is divided using a segmentation method to obtain the divided spatial units; Extract the stratigraphic information within each spatial unit. If there are missing or abnormal stratigraphic information, interpolate the stratigraphic information from neighboring spatial units to obtain the initial stratigraphic sequence.
[0009] Preferably, the method for extracting local geological features based on the initial stratigraphic sequence and obtaining thickness variation trend data based on the local geological features includes: Local geological features are extracted based on the initial stratigraphic sequence. Thickness differences between adjacent sequences are calculated, and the Kriging interpolation method is used to smooth these differences. A variogram model is constructed to analyze the spatial autocorrelation of thickness variations, determine the optimal interpolation parameters, and thus generate thickness variation trend data, specifically including: The variance of the variogram model is: ; in, c ( h () represents the interval distance. h The variance of the sample pairs, z nFor the first n The value of each sample In order to be with the first n Sample interval h The values of the samples at each distance, N The total number of samples; Further construct the spherical model: ; in, c For variance, c 0 represents distance h The variance when it is 0 a This is the critical distance value; Based on the spherical model, the weighted least squares method is used to analyze the spatial autocorrelation of thickness variation and determine the optimal interpolation parameters.
[0010] Preferably, the method for generating a local three-dimensional mesh based on thickness variation trend data to obtain a preliminary ore body interface includes: Based on thickness change trend data, the location of thickness abrupt changes is obtained through a preset abrupt change threshold; Using the locations of abrupt changes in thickness as constraint nodes, the initial stratigraphic sequence is topologically updated to obtain the updated stratigraphic sequence. Based on the updated layer sequence and combined with the updated data at the locations of thickness abrupt changes, a grid cell distribution model is established at the locations of thickness abrupt changes. Based on the grid cell distribution model, radial basis functions are used for surface reconstruction to generate a preliminary ore body interface.
[0011] Preferably, the method for generating a preliminary ore body interface based on a grid cell distribution model and using radial basis functions for surface reconstruction includes: Extracting the normal vectors of nodes in 3D space: ; in, n For local surfaces p The normal vector; d From the origin of the coordinate system to the local surface p distance, p i For the first i One sampling point, k The number of sampling points; The modeling region is filled with a cube according to the grid size determined by the grid cell distribution model. An octree data structure is used to store the three-dimensional spatial nodes and their corresponding normal vectors to establish an implicit field. ; in, ψ (·) represents the radial basis functions. Represents the gradient; xj ∈ R 3 For the modeling region, the first j A three-dimensional spatial node, n The number of spatial nodes; α i , β i These are the coefficients of the radial basis functions; The implicit field is rendered using the moving tetrahedron algorithm to generate a preliminary ore body interface.
[0012] Preferably, the method for obtaining the extension path by tracing the direction of the initial ore body interface to acquire the predicted point set of the unexplored area includes: On the initial ore body interface, the spatial change trend of the interface is tracked to obtain the direction vector; Use the direction vector to generate a sketch of the extended path; Extract the boundaries of unexplored areas from the extended path sketch; For areas with unknown boundaries, a random forest algorithm is used to obtain a set of predicted points; Construct extended paths based on predicted point sets.
[0013] Preferably, the method for connecting disconnected areas using a volumetric filling method based on the extended path to obtain a complete ore body structure includes: Spatial analysis methods are used to digitally describe the extension path and obtain its spatial morphological distribution characteristics; Based on the spatial morphological distribution characteristics, the distribution range of the disconnected area is obtained; If the distribution range of the disconnected area exceeds the preset threshold, the volume element filling method is used to fill the disconnected area layer by layer to obtain the preliminary ore body structure framework. The preliminary framework structure was refined and adjusted using morphological optimization methods to obtain the complete ore body structure.
[0014] Preferred methods for verifying the continuity of the dip direction based on the complete ore body structure, determining overlapping areas, and obtaining the final three-dimensional reconstruction model include: Feature extraction is performed on the complete ore body structure to obtain dip direction features; If the extracted tendency direction features are consistent within a preset threshold range, it is determined that the continuous characteristics meet the requirements, and a continuous description of the tendency direction is obtained. By describing the continuous flow of data, the distribution of overlapping regions is analyzed, and region identification technology is used to identify these overlapping regions. The overlapping areas are processed by Boolean union to obtain the final 3D reconstruction model.
[0015] The present invention also provides a mineral deposit spatial structure reconstruction system based on three-dimensional modeling. The system is used to implement the aforementioned method and includes: a data acquisition module, an extraction module, a generation module, an extension module, a filling module, and a construction module. The acquisition module is used to collect strike and dip information of ore bodies within the ore deposit. Based on the strike and dip information of the ore bodies, the ore deposit space is divided into a continuous sequence to obtain the initial stratigraphic sequence. The extraction module is used to extract local geological features based on the initial stratigraphic sequence, and obtain thickness variation trend data based on the local geological features; The generation module is used to generate a local three-dimensional mesh based on thickness variation trend data to obtain the preliminary ore body interface; The extension module is used to trace the direction of the initial ore body interface, obtain the predicted point set of the unexplored area, and obtain the extension path. The filling module is used to connect disconnected areas based on the extended path using the volume element filling method to obtain the complete ore body structure; The module is used to verify the continuity of the dip direction based on the complete ore body structure, determine overlapping areas, and obtain the final three-dimensional reconstruction model.
[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention provides a method and system for reconstructing the spatial structure of mineral deposits based on 3D modeling. Addressing the core business scenario of difficulties in determining continuity due to complex ore body spatial morphology, abrupt thickness changes, and fault displacement, it achieves accurate modeling through multi-step fusion technology. By combining data-driven approaches with algorithmic collaboration, this invention solves the modeling challenges posed by the complexity of ore body structures and data sparsity, significantly improving the accuracy and reliability of ore body morphology prediction and providing efficient support for mineral resource exploration. Attached Figure Description
[0017] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a schematic diagram of the process of reconstructing the spatial structure of a mineral deposit based on three-dimensional modeling, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the spatial structure reconstruction system for mineral deposits based on three-dimensional modeling, according to an embodiment of the present invention. Detailed Implementation
[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0020] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0021] Example 1 This invention provides a method for reconstructing the spatial structure of mineral deposits based on three-dimensional modeling, comprising: Collect strike and dip information of ore bodies within the ore deposit, and based on the strike and dip information of ore bodies, divide the ore deposit space into a continuous sequence to obtain the initial stratigraphic sequence; Local geological features are extracted based on the initial stratigraphic sequence, and thickness variation trend data are obtained based on the local geological features. Based on thickness variation trend data, a local three-dimensional mesh is generated to obtain the preliminary ore body interface; Based on the preliminary ore body interface tracing direction extension, the predicted point set of the unexplored area is obtained to obtain the extension path; Based on the extended path, the volume element filling method is used to connect the disconnected areas and obtain the complete ore body structure; Based on the verification of the continuity of the dip direction of the complete ore body structure, the overlapping areas are identified, and the final three-dimensional reconstruction model is obtained.
[0022] like Figure 1 As shown, the specific implementation process of the present invention is as follows: S1. Collect strike and dip information of the ore bodies within the deposit. Based on this information, segment the deposit space into a continuous sequence to obtain the initial stratigraphic sequence. Specifically: By collecting borehole and profile data of the deposit, extract the strike and dip information of the ore bodies. Based on this information, use a segmentation method to segment the deposit space into a continuous sequence to obtain the initial stratigraphic sequence. The specific process is as follows: By collecting borehole data and profile data of the ore deposit, the spatial positioning information of the ore body contained therein is extracted, including the strike (i.e. the direction of the ore body's extension on the horizontal plane, usually expressed as azimuth) and dip (i.e. the direction of the ore body's inclination in the vertical direction, usually expressed as dip direction and dip angle), thereby obtaining the strike and dip information of the ore body within the ore deposit.
[0023] Based on the strike and dip information of the ore body, combined with the spatial coordinates of the boreholes and inclination data, a preliminary spatial distribution record of the ore deposit is obtained.
[0024] The space of the ore deposit is discretized by segmentation method. Based on the continuity characteristics of the ore body in the preliminary spatial distribution record, the space of the ore deposit is divided into several spatial units. Each spatial unit has a relatively independent spatial boundary in geometry, while maintaining a certain continuity and comparability in geology.
[0025] Stratigraphic information within each spatial unit is extracted, including single-layer thickness, lithological assemblage, contact relationships, and their spatial configuration with the ore body. If the stratigraphic information of a certain spatial unit has missing data, insufficient sampling, or data anomalies, such as abrupt changes in thickness, then based on the principle of spatial proximity, interpolation is performed using the stratigraphic information of its neighboring spatial units to obtain the initial stratigraphic sequence. Interpolation methods can include inverse distance weighting, geostatistical methods, etc.
[0026] S2. Extract local geological features based on the initial stratigraphic sequence, and obtain thickness variation trend data based on these features. Specifically, extract local geological features from the initial stratigraphic sequence, calculate the thickness difference between adjacent sequences, smooth the thickness difference using the Kriging interpolation method, construct a variogram model, analyze the spatial autocorrelation of thickness variations, determine the optimal interpolation parameters, and thus generate thickness variation trend data. The specific process is as follows: Based on the initial stratigraphic sequence, local geological features within each spatial unit are further extracted, thickness variations between adjacent sequences are analyzed, thickness difference values are calculated, and the spatial location, directional changes, and rate of change of the thickness differences are recorded to form a thickness variation dataset.
[0027] Based on the thickness variation dataset, the Kriging interpolation method is used to smooth the thickness difference values between adjacent sequences. By constructing a variogram model, the spatial autocorrelation of thickness variation is analyzed, and the optimal interpolation parameters are determined, thereby generating continuous, smooth, and geologically significant thickness variation trend data.
[0028] Kriging interpolation expands the numerical values by utilizing the covariance between the sample point and the point / region / volume to be assigned, as well as the covariance of the sample points affecting that point. This invention defines a variogram model representing the relationship between distance and variance, where the variance is: in, c ( h () represents the interval distance. h The variance of the sample pairs, z n For the first n The value of each sample In order to be with the first n Sample interval h The values of the samples at each distance,N The total number of samples.
[0029] The further constructed spherical model can be represented as: in, c For variance, c 0 represents distance h The variance when it is 0 a This is the critical distance value.
[0030] Based on the spherical model, the weighted least squares method is used to analyze the spatial autocorrelation of thickness variation and determine the optimal interpolation parameters.
[0031] in, w n As weight, Z To be based on the surrounding N The values of the samples after smoothing using Kriging interpolation.
[0032] S3. Based on thickness variation trend data, a local 3D mesh is generated to obtain the preliminary ore body interface. Specifically: based on thickness variation trend data, the locations of thickness abrupt changes are obtained through a preset abrupt change threshold. Using these abrupt change locations as constraint nodes, the initial stratigraphic sequence is topologically updated. Combined with the updated data from these abrupt change locations, a local 3D mesh is generated. Radial basis functions are then used for surface reconstruction to obtain the preliminary ore body interface. The specific process is as follows: Based on thickness change trend data, the location of thickness abrupt changes is obtained through a preset abrupt change threshold.
[0033] Using thickness abrupt change locations as constraint nodes, the initial stratigraphic sequence is topologically updated: new control points are inserted at the thickness abrupt change locations, the stratigraphic surface is reparameterized, and the stratigraphic sequence is locally refitted using minimum curvature spline or thin plate spline interpolation methods to ensure that the surface remains continuous at the abrupt change locations and that the curvature extrema match the geological reality; at the same time, the correspondence and contact properties between stratigraphic sequences are updated to eliminate spurious transitions caused by sparse original sampling or smooth interpolation, and finally, an updated stratigraphic sequence that is geometrically accurate, geologically consistent, and contains abrupt change characteristics is obtained.
[0034] Based on the updated stratigraphic sequence and combined with the updated data at the thickness abrupt change locations, a high-resolution local 3D mesh is established at the thickness abrupt change locations: First, with the abrupt change line as the axis, an adaptive densification strategy using ellipsoids or cuboids is adopted to set the horizontal and vertical mesh sizes to generate a local mesh; then, the global stratigraphic surface is mapped to this mesh, and through truncation and attribute assignment, the lithology code, thickness, mineralization grade, etc., corresponding to each mesh unit are extracted to form a mesh unit distribution model that can be used for explicit surface reconstruction; Based on the grid cell distribution model, radial basis functions are used for surface reconstruction to generate a preliminary ore body interface: Radial basis function (HRBF) surfaces, constrained by the spatial coordinates of sampling points and their "inner / outer" attributes, linearly combine a series of radial basis functions into an implicit expression function. Calculating this function value on a regular spatial grid and extracting the zero isosurface yields a continuous three-dimensional surface. For subsurface sample data with spatial autocorrelation, the Hermite-type radial basis function (HRBF) utilizes the Euclidean distance between the node to be estimated and the sampling points while introducing the surface normal vector at each sampling point to jointly construct a locally supported and globally consistent implicit field. This strategy preserves all sampling information while suppressing long-distance interference through a tightly supported kernel function, maintaining both overall smoothness and local detail at complex orebody boundaries. Therefore, this invention selects HRBF as the core reconstruction algorithm.
[0035] HRBF performs three-dimensional ore body modeling in two steps: first, extract the ore body boundary and simultaneously estimate its normal vector; then, construct the implicit field of the ore body and visualize the implicit field, thus completing the spatial calculation and visualization of the implicit field and obtaining the three-dimensional ore body model.
[0036] HRBF distinguishes the location of 3D spatial nodes by calculating the function values of specific 3D spatial nodes: in, x i ∈ R 3 For the modeling region, the first i A three-dimensional spatial node.
[0037] Extracting the normal vectors of nodes in 3D space: in, n For local surfaces p The normal vector; d From the origin of the coordinate system to the local surface p distance, p i For the first i One sampling point, k This represents the number of sampling points.
[0038] Based on the preset level of detail in the ore body model, the modeling area is filled with cubes according to the grid size determined by the grid cell distribution model. An octree data structure is used to store the three-dimensional spatial nodes and their corresponding normal vectors. Hermite-type radial basis function coefficients are analyzed, and an implicit field is established. This invention selects a solid cube surface function as the basis function; therefore, the implicit field is: in, ψ (·) represents the radial basis functions. Represents the gradient; xj ∈ R 3 is the first in the modeling region j A three-dimensional spatial node, n The number of spatial nodes; αi , βi These are the coefficients of the radial basis functions.
[0039] At this point, the implicit field itself is only a scalar field and cannot directly represent the ore body shape. It must be translated into a renderable surface using a three-dimensional visualization algorithm. This invention uses the Marching Tetrahedra algorithm for rendering: first, the spatial cubic domain is subdivided into six tetrahedral units, and then the tetrahedron is divided into two sub-polygons within the unit according to the difference in the sign of the node field values. Different colors are assigned according to the node attributes for surface rendering. Finally, all zero isovalue surfaces are extracted and colored to achieve a three-dimensional display of the implicit surface, resulting in a preliminary ore body interface that can be used for subsequent analysis.
[0040] S4. Based on the preliminary ore body interface, trace the strike direction to extend the ore body and obtain the predicted point set for the unexplored area, thus obtaining the extension path. Specifically: On the preliminary ore body interface, trace the spatial change trend of the interface to obtain the strike direction vector, generate an extension path sketch, obtain the boundary of the unexplored area from it, use the random forest algorithm to obtain the predicted point set, and then construct the extension path. The specific process is as follows: On the preliminary orebody interface, the spatial variation trend of the interface is tracked to obtain the strike direction vector. The tracking results can be extracted from the features of a pre-set geological model, which typically contains spatial distribution information of the orebody. By analyzing the data geometric characteristics of the preliminary orebody interface, the outline of the orebody can be preliminarily delineated. Suppose that in a specific deposit, the data of the preliminary orebody interface indicates that the orebody exhibits a significant dip characteristic in a certain area, the tracking results may prioritize the spatial variation trend of this dip surface. Directional analysis is then performed on the spatial data from the tracking results. Suppose that in a certain tracking result, the orebody's strike shows a northeastward dip; then, by extracting the directional data from multiple interfaces within this area, an approximate strike direction vector can be calculated.
[0041] Using the strike direction vector, an extension path sketch is generated. When generating the extension path sketch, the strike direction vector can be used to simulate the potential extension trajectory of the ore body. In practice, the strike direction vector can be used as a baseline, combined with the interface data in the geological model, to draw a preliminary extension path. Assuming the strike direction of the ore body is 45 degrees northeast, a curve is drawn along this direction, covering the known interface data points, and extending into the unexplored area to generate the extension path sketch.
[0042] Extract the boundaries of unexplored areas from the extended path sketch. Spatially divide the areas not covered in the extended path sketch. Assuming the extended path sketch shows that the ore body extends for about 2 kilometers in the northeast direction before the data is missing, then this missing area can be regarded as the boundary of unexplored areas.
[0043] For unexplored region boundaries, a random forest algorithm is used to obtain a predicted point set. Based on known interface data and the characteristics of unexplored region boundaries, the random forest algorithm integrates multiple variables to generate a predicted point set.
[0044] An extension path is constructed based on the predicted point set. The predicted values in the predicted point set are arranged according to their spatial location to obtain a continuous coordinate sequence. The coordinate sequence is combined with the extension path sketch to form a complete extension path. If the complete extension path shows signs of sharp termination, the boundary constraints of the predicted point set are adjusted to obtain the final extension path.
[0045] S5. Based on the extended path, the volumetric filling method is used to connect the disconnected areas to obtain the complete ore body structure. Specifically, spatial analysis methods are used to digitally describe the extended path, obtaining its spatial morphological distribution characteristics, and then the distribution range of the disconnected areas is determined. If the range exceeds a preset threshold, the volumetric filling method is used to fill the disconnected areas layer by layer. Then, morphological optimization methods are used to further refine and adjust the structure, thereby obtaining the complete ore body structure. The specific process is as follows: Spatial analysis methods are used to digitally describe the extension path and obtain its spatial morphological distribution characteristics. Three-dimensional curve parametric techniques (such as B-splines or NURBS) are used to fit the extension path into a continuous and differentiable spatial curve, and multi-dimensional geometric indicators such as curvature, torsion, extension direction vector, and local thickness change rate are extracted to form the spatial morphological distribution characteristics.
[0046] Based on the spatial morphological distribution characteristics, the distribution range of the disconnected areas is obtained. A strategy combining sliding window anomaly detection and topological connectivity analysis is introduced: potential disconnected areas are automatically marked using criteria such as curvature abrupt changes, torsion sign reversal, and a sudden reduction in local thickness to below 30% of the background value; simultaneously, the α-shape algorithm is used to reconstruct the convex hull of the point cloud, and the spatial ellipsoidal parameters (major semi-axis a, middle semi-axis b, minor semi-axis c) of each disconnected area are calculated to accurately obtain its center coordinates, volume, maximum penetration distance, and minimum gap with the main ore body, thereby obtaining the distribution range of the disconnected areas.
[0047] If the distribution range of the disconnected area exceeds a preset threshold, the volumetric filling method is used to fill the disconnected area layer by layer to obtain a preliminary ore body structure framework. If the maximum penetration distance or volume of any disconnected area exceeds a preset threshold, the volumetric filling operation is triggered. The application of the volumetric filling method can be imagined as adding virtual volumes layer by layer within the disconnected area to fill data gaps. Assuming that a disconnected area is 20 meters deep and 10 meters wide, it can be filled step by step with a thickness of 1 meter per layer to ensure data continuity.
[0048] The initial framework structure is refined and adjusted using morphological optimization methods to obtain the complete ore body structure. For example, when refining the ore body structure, morphological optimization methods can adjust areas with uneven boundaries to make the boundary lines more natural.
[0049] S6. Based on the complete ore body structure, verify the continuity of the dip direction, identify overlapping areas, and obtain the final 3D reconstruction model. Specifically: extract features from the complete ore body structure to obtain dip direction features, verify the continuity of the dip direction, use region identification technology to identify overlapping areas, perform Boolean union processing on the overlapping areas, and obtain the final 3D reconstruction model. The specific process is as follows: Feature extraction is performed on the complete orebody structure to obtain dip direction features. A method based on local surface fitting is used to construct a least-squares plane at the vertex of each complete orebody structure, calculate its normal vector, and further convert it into dip direction (i.e., the projection direction of the normal vector on the horizontal plane, usually represented by azimuth).
[0050] If the extracted dip direction features show consistency within a preset threshold range, the continuity characteristic is determined to meet the requirements, and a continuity description of the dip direction is obtained. By setting a preset threshold range, the complete ore body structure is traversed and judged. If more than 85% of the nodes in a certain area meet the consistency condition, the continuity characteristic of that area is considered to meet the modeling requirements and is retained; otherwise, it is marked as a potential discontinuous area.
[0051] By analyzing the distribution of overlapping regions through continuous description, region identification technology is employed to identify these regions. Since ore bodies may exhibit complex geometric relationships such as branching, merging, or intersections in three-dimensional space, region identification technology is needed to identify and divide geometrically overlapping parts of the complete ore body structure that differ in attributes or origins. First, the complete ore body structure is voxelized to construct a three-dimensional binary raster. Then, a connected component analysis algorithm is used to identify all independent connected regions. Next, it is determined which regions have spatial overlap but conflicting attributes. Finally, the set of all regions satisfying the overlap condition is extracted, and their spatial extent, volume, origin, and degree of overlap are recorded to obtain the overlapping regions.
[0052] The overlapping regions are processed using Boolean union to obtain the final 3D reconstruction model. For the identified overlapping regions, Boolean union is used for geometric fusion to eliminate redundant representations and unify the structure: First, the triangular meshes corresponding to each overlapping region are imported into the Boolean operation engine, and a spatial segmentation tree (such as a BSP tree or OBB tree) is used to accelerate intersection detection and mesh reconstruction; then, the intersection boundaries are re-triangulated to ensure that the newly generated surfaces are geometrically seamless and topologically correct; finally, a 3D reconstruction model with no redundancy, no overlap, and consistent structure is output.
[0053] In summary, this invention discloses a method for reconstructing the spatial structure of ore deposits based on 3D modeling. Addressing the core business scenario problem of difficulties in determining continuity due to complex ore body spatial morphology, abrupt thickness changes, and fault displacement, it achieves accurate modeling through multi-step fusion technology. This invention, through data-driven and algorithmic collaboration, solves the modeling challenges posed by the complexity of ore body structures and data sparsity, significantly improving the accuracy and reliability of ore body morphology prediction and providing efficient support for mineral resource exploration.
[0054] Example 2 like Figure 2 As shown, based on the same inventive concept, the present invention also provides a mineral deposit spatial structure reconstruction system based on three-dimensional modeling, used to implement the method described in the foregoing embodiments. The system includes: a collection module, an extraction module, a generation module, an extension module, a filling module, and a construction module. The acquisition module is used to collect strike and dip information of ore bodies within the ore deposit. Based on the strike and dip information of the ore bodies, the ore deposit space is divided into a continuous sequence to obtain the initial stratigraphic sequence. The extraction module is used to extract local geological features based on the initial stratigraphic sequence, and obtain thickness variation trend data based on the local geological features; The generation module is used to generate a local three-dimensional mesh based on thickness variation trend data to obtain the preliminary ore body interface; The extension module is used to trace the direction of the initial ore body interface, obtain the predicted point set of the unexplored area, and obtain the extension path. The filling module is used to connect disconnected areas based on the extended path using the volume element filling method to obtain the complete ore body structure; The module is used to verify the continuity of the dip direction based on the complete ore body structure, determine overlapping areas, and obtain the final three-dimensional reconstruction model.
[0055] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.
Claims
1. A method for reconstructing spatial structure of a deposit based on three-dimensional modeling, characterized in that, The method includes: Collect strike and dip information of ore bodies within the ore deposit, and based on the strike and dip information of ore bodies, divide the ore deposit space into a continuous sequence to obtain the initial stratigraphic sequence; Local geological features are extracted based on the initial stratigraphic sequence, and thickness variation trend data are obtained based on the local geological features. Based on thickness variation trend data, a local three-dimensional mesh is generated to obtain the preliminary ore body interface; Based on the preliminary ore body interface tracing direction extension, the predicted point set of the unexplored area is obtained to obtain the extension path; Based on the extended path, the volume element filling method is used to connect the disconnected areas and obtain the complete ore body structure; Based on the verification of the continuity of the dip direction of the complete ore body structure, the overlapping areas are identified, and the final three-dimensional reconstruction model is obtained.
2. The method of claim 1, wherein, Methods for collecting strike and dip information of ore bodies within a ore deposit, and then segmenting the ore deposit space into a continuous sequence based on this information to obtain an initial stratigraphic sequence include: By collecting borehole data and profile data of the ore deposit, the strike and dip information of the ore body within the ore deposit can be obtained; Based on the strike and dip information of the ore body, a preliminary spatial distribution record of the ore deposit is obtained; Based on the preliminary spatial distribution record, the space of the ore deposit is divided using a segmentation method to obtain the divided spatial units; Extract the stratigraphic information within each spatial unit. If there are missing or abnormal stratigraphic information, interpolate the stratigraphic information from neighboring spatial units to obtain the initial stratigraphic sequence.
3. The method according to claim 1, characterized in that, Methods for extracting local geological features based on initial stratigraphic sequences and obtaining thickness variation trend data based on local geological features include: Local geological features are extracted based on the initial stratigraphic sequence. Thickness differences between adjacent sequences are calculated, and the Kriging interpolation method is used to smooth these differences. A variogram model is constructed to analyze the spatial autocorrelation of thickness variations, determine the optimal interpolation parameters, and thus generate thickness variation trend data, specifically including: The variance of the variogram model is: ； in, c ( h () represents the interval distance. h The variance of the sample pairs, z n For the first n The value of each sample In order to be with the first n Sample interval h The values of the samples at each distance, N The total number of samples; Further construct the spherical model: ； in, c For variance, c 0 represents distance h The variance when it is 0 a This is the critical distance value; Based on the spherical model, the weighted least squares method is used to analyze the spatial autocorrelation of thickness variation and determine the optimal interpolation parameters.
4. The method according to claim 1, characterized in that, Methods for generating local 3D meshes and obtaining preliminary orebody interfaces based on thickness variation trend data include: Based on thickness change trend data, the location of thickness abrupt changes is obtained through a preset abrupt change threshold; Using the locations of abrupt changes in thickness as constraint nodes, the initial stratigraphic sequence is topologically updated to obtain the updated stratigraphic sequence. Based on the updated layer sequence and combined with the updated data at the locations of thickness abrupt changes, a grid cell distribution model is established at the locations of thickness abrupt changes. Based on the grid cell distribution model, radial basis functions are used for surface reconstruction to generate a preliminary ore body interface.
5. The method according to claim 4, characterized in that, The methods for generating preliminary orebody interfaces based on a grid cell distribution model and using radial basis functions for surface reconstruction include: Extracting the normal vectors of nodes in 3D space: ； in, n For local surfaces p The normal vector; d From the origin of the coordinate system to the local surface p distance, p i For the first i One sampling point, k The number of sampling points; The modeling region is filled with a cube according to the grid size determined by the grid cell distribution model. An octree data structure is used to store the three-dimensional spatial nodes and their corresponding normal vectors to establish an implicit field. ； in, ψ (·) represents the radial basis functions. Represents the gradient; x j ∈ R 3 For the modeling region, the first j A three-dimensional spatial node, n The number of spatial nodes; α i , β i These are the coefficients of the radial basis functions; The implicit field is rendered using the moving tetrahedron algorithm to generate a preliminary ore body interface.
6. The method according to claim 1, characterized in that, Methods for obtaining the extension path by tracing the orientation of the preliminary ore body interface and acquiring the predicted point set of unexplored areas include: On the initial ore body interface, the spatial change trend of the interface is tracked to obtain the direction vector; Use the direction vector to generate a sketch of the extended path; Extract the boundaries of unexplored areas from the extended path sketch; For areas with unknown boundaries, a random forest algorithm is used to obtain a set of predicted points; Construct extended paths based on predicted point sets.
7. The method according to claim 1, characterized in that, Based on the extended path, the method of connecting the disconnected areas using the volume element filling method to obtain the complete ore body structure includes: Spatial analysis methods are used to digitally describe the extension path and obtain its spatial morphological distribution characteristics; Based on the spatial morphological distribution characteristics, the distribution range of the disconnected area is obtained; If the distribution range of the disconnected area exceeds the preset threshold, the volume element filling method is used to fill the disconnected area layer by layer to obtain the preliminary ore body structure framework. The preliminary framework structure was refined and adjusted using morphological optimization methods to obtain the complete ore body structure.
8. The method according to claim 1, characterized in that, The methods for verifying the continuity of the dip direction based on the complete ore body structure, identifying overlapping areas, and obtaining the final 3D reconstruction model include: Feature extraction is performed on the complete ore body structure to obtain dip direction features; If the extracted tendency direction features are consistent within a preset threshold range, it is determined that the continuous characteristics meet the requirements, and a continuous description of the tendency direction is obtained. By describing the continuous flow of data, the distribution of overlapping regions is analyzed, and region identification technology is used to identify these overlapping regions. The overlapping areas are processed by Boolean union to obtain the final 3D reconstruction model.
9. A mineral deposit spatial structure reconstruction system based on three-dimensional modeling, the system being used to implement the method described in any one of claims 1-8, characterized in that, The system includes: a data acquisition module, an extraction module, a generation module, an extension module, a filling module, and a construction module; The acquisition module is used to collect strike and dip information of ore bodies within the ore deposit. Based on the strike and dip information of the ore bodies, the ore deposit space is divided into a continuous sequence to obtain the initial stratigraphic sequence. The extraction module is used to extract local geological features based on the initial stratigraphic sequence, and obtain thickness variation trend data based on the local geological features; The generation module is used to generate a local three-dimensional mesh based on thickness variation trend data to obtain the preliminary ore body interface; The extension module is used to trace the direction of the initial ore body interface, obtain the predicted point set of the unexplored area, and obtain the extension path. The filling module is used to connect disconnected areas based on the extended path using the volume element filling method to obtain the complete ore body structure; The module is used to verify the continuity of the dip direction based on the complete ore body structure, determine overlapping areas, and obtain the final three-dimensional reconstruction model.