A method, system and device for three-dimensional modeling of ore bodies based on optimal mapping of contour line skeletons

Abstract

This invention discloses a method, system, and device for 3D modeling of ore bodies based on optimal mapping of contour line skeletons. The method includes: acquiring a set of approximately parallel ore body contour lines; extracting the internal and external skeletons of each contour line; classifying the skeletons and identifying primary and secondary feature points; constructing feature vectors for each feature point; calculating the matching cost of primary feature points, selecting the optimal starting point, and rearranging the sequence; using an improved dynamic programming algorithm to solve for the optimal mapping of feature points that satisfies one-to-one mapping, order preservation, and global optimality; segmenting the contour lines according to the optimal mapping and constructing a triangular mesh, and merging them to obtain a complete 3D model of the ore body. This invention achieves accurate and automatic feature point correspondence and high-quality 3D model construction between ore body contour lines through the fusion of internal and external skeletons, skeleton-level guided feature recognition, and optimal mapping with reward and penalty mechanisms, significantly improving modeling efficiency and geometric rationality.

Description

Technical Field

[0001] This invention relates to the field of 3D modeling of ore bodies, and specifically to a method, system, and device for 3D modeling of ore bodies based on optimal mapping of contour skeletons. Background Technology

[0002] In the fields of geological exploration and mineral resource assessment, three-dimensional geological modeling is a key technical means to accurately depict the spatial morphology of ore bodies, precisely estimate reserves, and optimize mining plans. The input data for three-dimensional ore body modeling is usually a set of approximately parallel ore body contour lines. These contour lines are closed polylines on a spatial plane. By constructing a triangular network between adjacent contour lines, a complete three-dimensional model of the ore body can be formed.

[0003] Traditional 3D modeling methods for ore bodies often rely on manual experience to determine the correspondence and connection of contour lines. Technicians need to manually determine the correspondence between feature points on adjacent contour lines and construct triangular mesh surfaces based on this. This method is not only inefficient but also makes it difficult to guarantee the geometric rationality and topological correctness of the model under complex morphological changes. Especially when dealing with ore body contour lines that vary in shape, size, and concavity / convexity, it is essential to establish the correct correspondence between feature points through human-computer interaction; otherwise, geometric errors such as triangular face distortion and self-intersection are prone to occur, seriously affecting the model quality.

[0004] To address the aforementioned issues, existing research has attempted to introduce central axes or skeleton lines to characterize the internal topological structure of two-dimensional graphics and use them as the basis for contour line correspondence analysis. Extracting skeleton information from contour lines can reflect the morphological characteristics of graphics to some extent, providing a reference for feature point matching. However, traditional skeleton extraction methods are sensitive to contour noise and typically only focus on extracting the internal skeleton, failing to effectively combine the external expansion morphological information of the graphics, resulting in an incomplete representation of contour line morphology. Furthermore, how to identify the dominant structural lines from complex skeleton networks and quantitatively evaluate their characteristics to establish robust point-to-point matching still lacks a systematic solution.

[0005] At the matching algorithm level, existing technologies mostly employ simple nearest neighbor matching or dynamic programming methods. While these methods can guarantee the order of feature points to some extent, they struggle to flexibly handle situations where the number of feature points on two contour lines is inconsistent. When the number of feature points on two contour lines differs significantly, it can lead to a large number of feature points being forcibly matched or ignored, resulting in the loss of important morphological details or the introduction of incorrect connections. Achieving a balance between matching quality and continuity, and realizing an optimal mapping that maintains both topological coherence and geometric similarity, is a significant challenge currently facing these technologies. Summary of the Invention

[0006] To address the problems existing in the prior art, the present invention aims to provide a method, system, and device for 3D modeling of ore bodies based on optimal mapping of contour skeletons. By integrating the inner and outer skeletons to fully represent the contour morphology, and using the main skeleton to guide feature recognition, the rationality and geological consistency of point-to-point mapping under complex morphological changes are ensured. This effectively avoids geometric errors such as triangular face distortion and self-intersection common in traditional methods, and significantly improves the quality of model construction.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A 3D modeling method for ore bodies based on optimal mapping of contour skeletons is proposed. The method inputs a set of approximately parallel ore body contour lines, each contour line being a closed polyline on a spatial plane. Adjacent contour lines are connected by a wireframe to construct a triangular mesh, thereby obtaining a 3D model of the ore body. The method specifically includes the following steps: Step 1: For the outline of adjacent ore bodies and The internal and external skeletons of each ore body contour line were extracted using the wavefront contraction method. Step 2: Classify the extracted skeleton and identify the main and secondary feature points of the ore body outline, collectively referred to as ore body outline feature points; Step 3: Construct feature vectors for each ore body contour feature point. The feature vectors include the length of the main skeleton, the orientation of the main skeleton, the average length of the supporting skeleton, and the local curvature. Step 4: Calculate the matching cost between the main feature points of two adjacent ore body contour lines. The matching cost integrates spatial distance, difference in main skeleton length, difference in main skeleton orientation, difference in average length of supporting skeleton, and difference in local curvature. Select the main feature point pair with the minimum matching cost as the starting point of the feature point sequence, and rearrange the feature point sequence to obtain the ore body contour line. The feature point sequence is ore body outline The feature point sequence is ; Step 5: Use dynamic programming algorithm to solve for the optimal mapping between the feature point sequences of two adjacent ore body contour lines. The optimal mapping satisfies one-to-one mapping, order preservation and global optimality. Step 6: Based on the optimal mapping, decompose the two adjacent ore body contour lines into multiple non-closed curve segments. Construct a triangular mesh for each pair of corresponding curve segments, merge all triangular meshes to obtain the triangular mesh surface between adjacent contour lines, and merge the triangular mesh surfaces of all adjacent contour line pairs to obtain the complete ore body 3D model.

[0008] Step 1 is described in detail as follows: For the outline of adjacent ore bodies and ore body outline and Perform direction consistency processing; Then, the contour lines of the ore bodies were extracted using the wavefront contraction method. and Internal skeleton and The inner normal direction of the wavefront contour line contracts uniformly inward, recording the motion trajectory of all vertices during the wavefront motion, and the union of all trajectories constitutes the internal skeleton. Next, the contour lines of the ore bodies were extracted using the wavefront contraction method. and external skeleton and First, calculate the convex hull of the contour line. Adjust the wavefront motion direction to the outward normal direction. Expand outward with the convex hull boundary as the termination boundary. Record the motion trajectory of all vertices. Their union constitutes the external skeleton. Step 2 is described in detail below: First, the skeleton lines that intersect with the ore body outline or convex hull boundary are defined as the supporting skeleton lines, and the skeleton lines that do not intersect with the ore body outline or convex hull boundary are defined as the main skeleton lines. The internal skeleton of the ore body outline is regarded as an acyclic connected undirected graph. The longest path of the internal skeleton of the ore body outline is obtained by using a depth-first search or breadth-first search algorithm. This longest path is defined as the main skeleton, and the other paths of the internal skeleton of the ore body outline and the paths of the external skeleton are defined as the secondary skeleton.

[0009] Then, the branches connected to the endpoints of the main skeleton or secondary skeleton are traversed sequentially, and the longest branch connected to each endpoint is merged into the main skeleton or secondary skeleton. The endpoints of the main skeleton are defined as primary feature points, and the endpoints of the secondary skeleton are defined as secondary feature points. The primary and secondary feature points are collectively referred to as ore body contour feature points.

[0010] In step 3, for each ore body contour feature point The spatial coordinates of the point Its eigenvectors are calculated as follows: skeleton length Feature points The associated skeleton length; Skeletal position Feature points The orientation of the associated skeleton is defined as a vector. The azimuth angle; Average length of the support frame Feature points The average length of the adjacent branch skeletons of the main skeleton; Local features Feature points Local contextual features along the ore body outline are obtained through discrete curvature approximation; let feature points be... The adjacent points on the outline are respectively and Their coordinates are ( )and( If ), then the local feature is defined as: ,in, , , , .

[0011] In step 4, the matching cost is calculated as follows:

[0012]

[0013] In the formula, Represents the outline of the ore body Main feature points Outline of the ore body Main feature points The matching cost, , , and These represent the weights of the trunk skeleton length, trunk skeleton orientation, average length of the branch skeleton, and local features, respectively.

[0014] In step 5 One-to-one mapping: π is a one-to-one mapping, that is, for any There exists a unique one. Make And inverse mapping The presence of a feature point indicates that each selected feature point can be matched at most once. Order preservation: if ( i <i' ),but ; indicates if exist Middle row Previously, they were in The corresponding points in the text must also maintain the same order: exist Middle row Before; Optimality: The sum of the matching costs of all matching pairs is minimized globally.

[0015] The dynamic programming algorithm in step 5 is specifically as follows: (1) State definition main state : forward points and forward Minimum total matching cost for points; Auxiliary status :arrive( , )hour The number of consecutive skipped points; Auxiliary status :arrive( , )hour The number of consecutive skipped points; Auxiliary status :arrive( , The number of point pairs matched when ).

[0016] (2) Boundary condition initialization For all Initialize the main state Auxiliary state Auxiliary state Auxiliary state ; For the first column: , , ; For the first line: , , ; in, This indicates that the basic penalty is skipped. The mean of the cost matrix, This is an adjustment factor, ranging from 0.1 to 0.5; (3) State transition equation for All other states will be calculated step by step during the state transition process, and the calculation equations are as follows:

[0017] In the formula, under the matching operation, feature points and Establish a correspondence; from the state ( The total cost increases due to the transfer. ,in, This is the consecutive match reward coefficient; after a match, the consecutive skip count is reset. , ; Number of matching pairs ; jump over Feature points in This indicates that feature points will not be included. If it matches any point, skip it directly; from the state ( The total cost increases due to the transfer. ,in, It continuously skips increasing coefficients; it skips feature points. back, , ; jump over Feature points in This indicates that feature points will not be included. If it matches any point, skip it directly; from the state ( The total cost increases due to the transfer. Skip feature points back, , ; Through the above state transition operations, a feature point mapping relationship that satisfies one-to-one mapping, order preservation, and optimality is finally obtained. .

[0018] A 3D modeling system for ore bodies based on optimal mapping of contour skeletons, comprising: The data input module is used to input a set of approximately parallel ore body contour lines. Each ore body contour line is a closed polyline on a spatial plane, consisting of a series of vertices arranged in sequence; The skeleton extraction module is used to ensure the directionality of adjacent contour lines and to extract the internal and external skeletons of each ore body contour line; this skeleton extraction module specifically includes: Direction unification unit: For a pair of adjacent ore body contour lines currently being processed, the vertex sequences of the two are processed to unify the direction to either clockwise or counterclockwise, so that the normals of the two contour lines are consistent. Internal skeleton extraction unit: The wavefront contraction method is adopted, and the inner normal direction of the wavefront contour line is uniformly contracted inward. The motion trajectory of all vertices during the wavefront motion is recorded, and the union of all trajectories constitutes the internal skeleton.

[0019] External skeleton extraction unit: First, calculate the convex hull of the contour line, and then use the same wavefront contraction principle as the internal skeleton, but adjust the wavefront motion direction to the outward normal direction. Expand outward with the convex hull boundary as the termination boundary, record the motion trajectory of all vertices, and their union constitutes the external skeleton. The skeleton hierarchy module is used to identify dominant feature points from the skeleton network. This skeleton hierarchy module includes: Skeleton classification unit: The skeleton line that intersects with the ore body outline or convex hull boundary is defined as the branch skeleton, and the skeleton line that does not intersect with the ore body outline or convex hull boundary is the main skeleton. Longest path search unit: The internal skeleton of the ore body outline is regarded as an acyclic connected undirected graph. The longest path is extracted as the main skeleton using depth-first search or breadth-first search algorithm. Other paths of the internal skeleton of the ore body outline and the paths of the external skeleton are used as secondary skeletons. Support skeleton merging unit: Traverse the support skeletons connected to the endpoints of the main skeleton or secondary skeleton, and merge the longest support skeleton connected to each endpoint into the main skeleton or secondary skeleton, so that the endpoints of the main skeleton fall on the ore body outline or convex hull boundary. Feature point definition unit: The endpoints of the main skeleton are defined as the main feature points of the ore body outline, and the endpoints of the secondary skeleton are defined as the secondary feature points of the ore body outline. The main and secondary feature points of the ore body outline are collectively referred to as the feature points of the ore body outline. The feature vector construction module is used to construct a four-dimensional feature vector for each feature point, quantifying its geometric and topological properties. The four-dimensional feature vector includes: Skeleton length: The spatial distance from a feature point along the skeleton to its opposite endpoint; Skeleton orientation: The orientation angle of the skeleton associated with the feature point; Average length of the scaffold: The average length of the scaffold adjacent to the main scaffold containing the feature point. Local curvature: an approximation of the discrete curvature of a feature point on a contour line; The starting point selection module is used to determine the optimal matching starting point for the feature point sequences of two contour lines. Specifically, the starting point selection module calculates the matching cost between all pairs of main feature points. The matching cost integrates spatial distance, difference in backbone length, difference in backbone orientation, difference in average length of branches, and difference in local curvature. The main feature point pair with the minimum cost is selected as the serialization starting point. The two feature point sequences are rearranged to align the starting points, resulting in the feature point sequences of the two ore body contour lines. and ; The optimal mapping module employs an improved dynamic programming algorithm to solve for the optimal mapping of feature points along the ore body contour, while satisfying one-to-one mapping, order preservation, and global optimality. This is expressed as: .

[0020] The model building and output module is used to construct a 3D surface based on the optimal mapping results and output the final model. Specifically, the model building and output module includes: Segmented Unit: Based on the matching feature point pairs in the optimal mapping, the two contour lines are decomposed into multiple non-closed curve segments respectively; Triangular mesh construction unit: For each pair of corresponding curve segments, a triangular mesh is constructed using the synchronous forward method, the minimum area method, or the minimum perimeter method; Merging Units: Merge all segmented triangular meshes to obtain a complete triangular mesh surface between adjacent contour lines; Output Unit: Merges the triangular mesh surfaces of all adjacent ore body outline pairs to form a complete 3D model of the ore body and outputs it.

[0021] In the process of solving the optimal mapping of feature points of the ore body contour line by the optimal mapping module, the one-to-one mapping property requires that the mapping is one-to-one, that is, each selected feature point can be matched at most once, and the number of matched point pairs is equal; the order preservation property requires that the mapping does not destroy the topological order of the contour line itself; the optimality property requires that the sum of the matching costs of all matched point pairs reaches the global minimum. The optimal mapping module uses a dynamic programming algorithm to solve for the optimal mapping of feature points on the ore body contour line. Specifically, it first initializes the boundary states and handles the case where only one sequence contains feature points, accumulating skip penalties sequentially. Then, it calculates each state in ascending order and selects the transition path with the minimum total cost from three operations: matching the current point, skipping the current point of the first sequence, and skipping the current point of the second sequence. Among these operations, the matching operation introduces a continuous matching reward coefficient to encourage the formation of continuous correspondences; the skipping operation introduces an increasing penalty for continuous skipping, so that the more consecutive skips there are, the heavier the penalty becomes, thereby suppressing unreasonable long-segment skips.

[0022] An electronic device for 3D modeling of ore bodies that takes into account both contour line geometric feature matching, comprising: At least one processor, The memory that is communicatively connected to the at least one processor The user interface is used to receive ore body contour data or modeling parameters input by the user. A network interface is used to obtain ore body contour data from an external database or output a 3D ore body model. The memory stores a computer program that can be executed by the at least one processor. When the computer program is executed by the processor, it causes the processor to perform a 3D modeling method for ore bodies based on optimal mapping of contour skeletons as described above.

[0023] By adopting the above scheme, this invention fully characterizes the morphological features of the contour lines by integrating the internal and external skeletons, overcoming the shortcomings of traditional methods that only focus on the internal skeleton while ignoring external expansion morphological information. Based on this, through skeleton hierarchical processing, the main and secondary skeletons are identified from the complex skeleton network, guiding the stable extraction of key feature points. The improved dynamic programming algorithm introduces a continuous matching reward mechanism during feature point matching, encouraging the formation of coherent correspondences and avoiding discontinuous matching results; simultaneously, a continuous skipping incremental penalty mechanism effectively suppresses unreasonable long-segment skips, preventing the loss of important morphological details. The comprehensive application of the above techniques ensures the rationality and geological consistency of point-to-point mapping under complex morphological changes, effectively avoiding geometric errors such as triangular face distortion and self-intersection common in traditional methods, and significantly improving the construction quality of the 3D model. Attached Figure Description

[0024] Picture 1 This is a flowchart of the method of the present invention; Picture 2 This is a system block diagram of the present invention; Picture 3 This is a block diagram of an electronic device. Detailed Implementation

[0025] like Picture 1 As shown, this invention discloses a 3D modeling method for ore bodies based on optimal mapping of contour skeletons. The method inputs a set of approximately parallel ore body contour lines, each contour line being a closed polyline on a spatial plane. Adjacent contour lines are connected by a wireframe to construct a triangular mesh, thereby obtaining a 3D model of the ore body. Specifically, the method includes the following steps: Step 1: Extract the inner and outer skeleton of the ore body outline.

[0026] For the outline of adjacent ore bodies and ore body outline and Perform directional alignment processing to make the ore body outline consistent. and The normal direction is consistent.

[0027] Then, the contour lines of the ore bodies were extracted using the wavefront contraction method. and Internal skeleton and The inner normal direction of the wavefront contour line contracts uniformly inward, recording the motion trajectory of all vertices during the wavefront motion. The union of all trajectories constitutes the internal skeleton.

[0028] Next, the contour lines of the ore bodies were extracted using the wavefront contraction method. and external skeleton and First, calculate the convex hull of the contour. Adjust the wavefront motion direction to the outward normal direction. Expand outward with the convex hull boundary as the termination boundary. Record the motion trajectory of all vertices. Their union constitutes the external skeleton.

[0029] Based on the outline of the ore body For example, the internal skeleton calculation and external skeleton calculation are as follows: 1. Internal skeleton calculation (1) Initialization of the wavefront system Outline Every single The original edge generates a corresponding wavefront. At the initial time t=0, we have All wavefronts are given the same constant velocity w and begin to translate along their inner normal direction (i.e., the direction pointing towards the interior region of the polygon). This setup constitutes an isotropic, uniformly contracting system.

[0030] (2) Vertex kinematics and trajectory tracking During wavefront motion, polygons each vertex It becomes a dynamic particle. Its trajectory is strictly determined by geometric constraints: vertex It always moves along the angle bisector of the angle formed by its two adjacent sides. Its instantaneous speed of motion... It is an included angle The function satisfies the relation The algorithm continues to solve. The parametric equations are used to accurately record its spatiotemporal trajectory. ,in This is the moment when the vertex disappears due to the event.

[0031] (3) Event-driven topology evolution The system's evolution is driven by a series of discrete topological events, which occur in the following three cases: Edge event: When the length of a wavefront segment is reduced to zero due to the movement of its two vertices towards each other, the segment disappears.

[0032] Split event: When a moving vertex The trajectory of the wavefront is related to another non-adjacent wavefront segment. When they intersect, the wavefront region is split at the intersection point.

[0033] Vertex event: The trajectories of two moving vertices intersect for the first time.

[0034] Let the time of occurrence of each event be . .exist At a certain moment, the topology of the wavefront network undergoes a transient abrupt change: edges disappear, regions split, or merge. After the event, the system continues to evolve under the new wavefront configuration, and the trajectories of the affected vertices and the timing of future potential events must be recalculated. The event queue is processed chronologically until all wavefronts have completely shrunk.

[0035] (4) Geometric construction of the skeleton Ore body outline Internal skeleton That is, the trajectory of all vertices Union of: .

[0036] 2. External skeleton calculation (1) Calculate the outline The convex hull is used as the boundary constraint for extracting the external skeleton. The convex hull is the smallest convex polygon that contains all vertices of the contour line, and the wavefront will terminate at the boundary of this convex hull during its outward expansion.

[0037] (2) Initialization of the wavefront system Outline Each original edge generates a corresponding wavefront. At the initial moment, the wavefront coincides with the original edge. All wavefronts are given the same constant velocity and translate along their respective outward normal directions (i.e., directions pointing towards the outer region of the polygon), forming an isotropic uniform expansion system.

[0038] (3) Vertex kinematics and trajectory tracking During wavefront motion, polygons each vertex It becomes a dynamic particle. Its trajectory is strictly determined by geometric constraints: vertex It always moves along the angle bisector of the angle formed by its two adjacent sides. Its instantaneous speed of motion... It is an included angle The function satisfies the relation The algorithm continuously solves the parametric equation, accurately recording the complete spatiotemporal trajectory of each vertex from the initial moment until it disappears due to a topological event or reaches the convex hull boundary.

[0039] (4) Event-driven topology evolution The system's evolution is driven by a series of discrete topological events, the same types as those extracted from the internal skeleton, including edge events, split events, and vertex events. At the moment each event occurs, the wavefront network's topology undergoes a transient abrupt change. After event processing, the system continues to evolve under the new wavefront configuration, recalculating the trajectories of affected vertices and the timing of future potential events. The event queue is processed chronologically until all wavefronts reach the convex hull boundary or disappear completely.

[0040] (5) Geometric construction of the skeleton Outline external skeleton That is, the union of the trajectories of all vertices.

[0041] This invention not only uses the wavefront contraction method to calculate its internal skeleton, but also further calculates its convex hull, and uses the same principle to contract towards the convex hull boundary to extract its external skeleton, thereby forming a complete internal and external topological representation of the contour morphology.

[0042] Step 2: Classification of ore body outline skeleton.

[0043] After extracting the inner and outer skeleton of the contour line, this embodiment further performs hierarchical processing on the skeleton to identify the feature points with dominant significance.

[0044] First, the skeleton lines that intersect with the ore body outline or convex hull boundary are defined as the branch skeleton, and the skeleton lines that do not intersect with the ore body outline or convex hull boundary are defined as the main skeleton.

[0045] The internal skeleton of the ore body outline is regarded as an acyclic connected undirected graph. The longest path of the internal skeleton of the ore body outline is obtained by using a depth-first search or breadth-first search algorithm. This longest path is defined as the main skeleton, and the other paths of the internal skeleton of the ore body outline and the paths of the external skeleton are defined as the secondary skeleton.

[0046] Then, the sub-skeletons connected to the endpoints of the main or secondary skeleton are traversed sequentially, and the longest sub-skeleton connected to each endpoint is merged into the main or secondary skeleton, so that the endpoints of the main or secondary skeleton are all on the ore body outline or bulge boundary. This merging operation ensures that the endpoints of all skeletons can establish a connection with the outline boundary.

[0047] The endpoints of the main skeleton are defined as the main feature points of the ore body outline, and the endpoints of the secondary skeleton are defined as the secondary feature points of the ore body outline. Both the main feature points and the secondary feature points of the ore body outline are called feature points of the ore body outline.

[0048] The above-mentioned hierarchical method can automatically identify feature points representing key parts of the contour shape from the complex skeleton network, providing stable anchor points for subsequent matching.

[0049] This invention defines "branch skeletons" and "trunk skeletons" on the generated skeleton network, and treats the internal trunk skeletons as acyclic graphs. It extracts the "main trunk skeleton" and "secondary trunk skeleton" through longest path search. Then, by traversing and merging the longest branch skeletons according to a rule, the endpoints of the trunk skeletons are uniformly placed on the original contour line or its convex hull boundary, thereby stably identifying the "primary feature points" and "secondary feature points".

[0050] Step 3: Construct the feature point vector of the ore body outline.

[0051] For each feature point on the ore body contour line, a four-dimensional feature vector is constructed to quantify the geometric and topological properties of that point. Let any feature point on the contour line... The spatial coordinates of the point The other end of its associated skeleton is This feature point The feature vectors include: (1) Length of the skeleton Feature points The associated backbone length, i.e., feature points The spatial distance along the trunk skeleton to its opposite end point. The length of the trunk skeleton. It reflects the scale of the skeleton branches represented by the feature points.

[0052] (2) Orientation of the skeleton Feature points The orientation of the associated skeleton is defined as a vector. The azimuth angle can be characterized by the angle between its horizontal projection and true north, or by the cosine of the three-dimensional direction. This feature reflects the direction of the skeleton's extension.

[0053] (3) Average length of the support frame Feature points The average length of the adjacent branches of the main skeleton. This feature reflects the density of skeletal branches around the feature point.

[0054] (4) Local features Feature points Local contextual features along the ore body outline are approximated using discrete curvature. Let feature points be... The adjacent points on the outline are respectively and Their coordinates are ( )and( If the local curvature is defined as follows: This feature reflects the degree of curvature of the contour line at that point.

[0055] in, , , , .

[0056] The above feature points If it is the outline of the ore body Feature points on, then = The above feature points If it is the outline of the ore body Feature points on, then = .

[0057] Based on the above definition, each feature point obtains a four-dimensional feature vector that integrates global topological attributes (trunk skeleton length, orientation, and average length of branch skeletons) and local geometric attributes (local curvature), providing a unified quantitative basis for subsequent feature point matching.

[0058] Step 4: Select the starting point for serializing the feature points of the ore body outline.

[0059] After constructing the feature vectors, it is necessary to determine the matching starting point of the feature point sequences of the two ore body contours in order to avoid mismatch due to improper selection of the starting point.

[0060] set up Represents the outline of the ore body Main feature points Outline of the ore body Main feature points The matching cost function.

[0061]

[0062]

[0063] In the formula, , , and These represent the weights of the trunk skeleton length, trunk skeleton orientation, average length of the branch skeleton, and local features, respectively, and can be adjusted according to the actual application scenario.

[0064] Calculate the matching cost between all pairs of principal feature points. The pair of principal feature points corresponding to the minimum value is the starting point for the serialization of feature points of the ore body contour line, thus obtaining the ore body contour line. The feature point sequence is ore body outline The feature point sequence is .

[0065] This optimization process effectively avoids matching deviations caused by improper selection of the starting point.

[0066] Step 5: Optimal Mapping of Feature Points on the Ore Body Outline: Using an improved dynamic programming algorithm, the optimal mapping of feature points on the ore body outline is solved while satisfying one-to-one mapping, order preservation, and global optimality. This is expressed as: .

[0067] One-to-one mapping: π is a one-to-one mapping, that is, for any There exists a unique one. Make And inverse mapping The existence of this feature point means that each selected feature point can be matched at most once. Order preservation: if ( i <i' ),but ; indicates if exist Middle row Previously (i.e.) i <i' ), then they are in The corresponding points in the text must also maintain the same order: exist Middle row Previously, this ensured that the mapping did not disrupt the topological order of the contour lines themselves.

[0068] Optimality: The sum of the matching costs for all matching pairs is minimized globally, i.e. Global minimum. Among them, Right now The comprehensive matching cost defined in step four integrates the differences in five aspects: spatial distance, trunk skeleton length, trunk skeleton orientation, average length of branch skeleton, and local curvature.

[0069] The process of solving for the optimal mapping of feature points of the ore body contour line using an improved dynamic programming algorithm is as follows: (1) State definition main state : forward points and forward Minimum total matching cost for points; Auxiliary status :arrive( , )hour The number of consecutive skipped points; Auxiliary status :arrive( , )hour The number of consecutive skipped points; Auxiliary status :arrive( , The number of point pairs matched when ).

[0070] (2) Boundary condition initialization For all Initialize the main state Auxiliary state Auxiliary state Auxiliary state ; For the first column: , , ; For the first line: , , ; in, This indicates that the basic penalty is skipped. The mean of the cost matrix, The adjustment factor is set to a value between 0.1 and 0.5.

[0071] (3) State transition equation All other states ( The calculation will be performed step by step during the state transition. The core logic of the state transition is to penalize skipping (especially consecutive skipping), incentivize continuous matching, and strictly maintain the order, as shown in the following equation:

[0072] In the formula, under the matching operation, feature points and Establish a correspondence. From the state ( The total cost increases due to the transfer. ,in, This is the consecutive matching reward coefficient (less than 1, typically 0.8-0.9), designed to encourage consecutive matching and reduce its cost. After a match, the consecutive skip count is reset. , ; Number of matching pairs .

[0073] jump over Feature points in This indicates that feature points will not be included. If a match is found at any point, skip it. From the state ( The total cost increases due to the transfer. ,in, This is the increment factor for consecutive skips (typically between 0.05 and 0.1). This design ensures that the more consecutive skips occur, the heavier the penalty for a single skip (the increment factor is 1 + μ × number of consecutive skips), thus suppressing unreasonable long skips. Skip characteristic points back, (Skip continuously) (increase the count by 1). (Skip continuously) (The count is reset to zero).

[0074] jump over Feature points in This indicates that feature points will not be included. If a match is found at any point, skip it. From the state ( The total cost increases due to the transfer. Skip feature points back, , .

[0075] Through the above state transition operations, a feature point mapping relationship that satisfies one-to-one mapping, order preservation, and optimality is finally obtained. .

[0076] This invention proposes an improved dynamic programming algorithm for finding the optimal bijective mapping between two contour feature point sequences. Its state transition equation not only calculates the matching cost but also introduces a continuous matching reward coefficient to encourage continuous correspondences and designs an incremental penalty term for continuous skipping, where μ acts on the number of consecutively skipped points s to adaptively penalize unreasonable skips. Thus, while maintaining the order, it intelligently handles the case of misaligned feature point numbers, obtaining a globally optimal and geologically reasonable mapping relationship.

[0077] Step 6: Construct a 3D model of the ore body by connecting the segments of the ore body outline.

[0078] To obtain the optimal mapping Then, sequentially from the mapping Extract the point pairs from the middle and outline the ore body. and It is decomposed into multiple segments. Each segment is a non-closed line, and its endpoints are determined by matching feature point pairs.

[0079] For each pair of adjacent feature points, the contour segments are connected using a classic triangulation algorithm. Optional methods include, but are not limited to, the synchronous forward method, the minimum area method, and the minimum perimeter method.

[0080] In this embodiment, the above methods can be flexibly selected according to the complexity of the contour lines. For contour lines with relatively regular shapes, the synchronous advancement method can achieve good results; for contour lines with complex shapes, the minimum area method or the minimum perimeter method can achieve better geometric rationality.

[0081] Finally, all the segmented triangular meshes are merged to obtain the ore body outline. and The complete triangular mesh surface between them.

[0082] Repeat steps 1-6 above to process all adjacent ore body contour pairs. After merging the triangular mesh surfaces of all adjacent contour pairs, a complete 3D model of the ore body is obtained. This invention achieves precise and automatic feature point mapping and triangulation construction between ore body contour lines through systematic skeleton extraction, grading, and feature quantization. The significant technical effects are reflected in three aspects: (1) Significant improvement in model construction quality: By integrating the inner and outer skeletons to fully represent the contour morphology, and using the main skeleton to guide feature recognition, the rationality and geological consistency of point-to-point mapping under complex morphological changes are ensured, and geometric errors such as triangular face distortion and self-intersection commonly found in traditional methods are effectively avoided. (2) The level of automation and intelligence in modeling has been greatly improved. The entire process from skeleton extraction to optimal mapping does not require manual intervention. The optimal mapping matching algorithm can effectively handle the situation where the number of feature points is inconsistent. While maintaining the order, it optimizes the matching continuity and global cost, and improves the robustness of handling complex and irregular contour lines. (3) Optimization of algorithm efficiency and practicality: The event-driven wavefront method efficiently extracts the skeleton, and the construction of feature vectors takes into account both global and local attributes. The matching algorithm balances detail preservation and computational efficiency through a continuous skip penalty mechanism, making the method applicable to large-scale and multi-form ore body modeling tasks in actual exploration, and providing a more reliable three-dimensional geological model basis for resource assessment and mining design.

[0083] Based on the same inventive concept, this invention also discloses a 3D modeling system for ore bodies based on optimal mapping of contour skeletons, which includes a data input module, a skeleton extraction module, a skeleton grading module, a feature vector construction module, a starting point selection module, an optimal mapping module, and a model construction and output module.

[0084] The data input module is used to input a set of approximately parallel ore body contour lines. Each ore body contour line is a closed polyline on a spatial plane, consisting of a series of vertices arranged in sequence.

[0085] The skeleton extraction module is used to ensure the orientation of adjacent contour lines and to extract the internal and external skeletons of each ore body contour line. Specifically, this skeleton extraction module includes: Orientation unification unit: For a pair of adjacent contour lines being processed, the vertex sequences of the two lines are unified into clockwise or counterclockwise directions to make the normals of the two contour lines consistent, laying the foundation for subsequent skeleton extraction and feature matching.

[0086] Internal skeleton extraction unit: The wavefront contraction method is adopted, and the inner normal direction of the wavefront contour line is uniformly contracted inward. The motion trajectory of all vertices during the wavefront motion is recorded, and the union of all trajectories constitutes the internal skeleton.

[0087] External skeleton extraction unit: First, calculate the convex hull of the contour line, and then use the same wavefront contraction principle as the internal skeleton, but adjust the wavefront motion direction to the outward normal direction. Expand outward with the convex hull boundary as the termination boundary, record the motion trajectory of all vertices, and their union constitutes the external skeleton.

[0088] This skeleton extraction module achieves a complete topological representation of the contour morphology by combining the internal and external skeletons.

[0089] The skeleton hierarchy module is used to identify dominant feature points from the skeleton network. This skeleton hierarchy module includes: Skeleton classification unit: Skeleton lines that intersect with the ore body outline or convex hull boundary are defined as branch skeletons, and skeleton lines that do not intersect with the ore body outline or convex hull boundary are defined as main skeletons.

[0090] Longest path search unit: The internal skeleton of the ore body outline is regarded as an acyclic connected undirected graph. The longest path is extracted as the main skeleton using a depth-first search or breadth-first search algorithm. Other paths of the internal skeleton of the ore body outline and the paths of the external skeleton are used as secondary skeletons.

[0091] Support skeleton merging unit: Traverse the support skeletons connected to the endpoints of the main skeleton or secondary skeleton, and merge the longest support skeleton connected to each endpoint into the main skeleton or secondary skeleton, so that the endpoints of the main skeleton fall on the ore body outline or convex boundary.

[0092] Feature point definition unit: The endpoints of the main skeleton are defined as the main feature points of the ore body outline, and the endpoints of the secondary skeleton are defined as the secondary feature points of the ore body outline. The main and secondary feature points of the ore body outline are collectively referred to as the feature points of the ore body outline.

[0093] The feature vector construction module is used to construct a four-dimensional feature vector for each feature point, quantifying its geometric and topological properties. The four-dimensional feature vector includes: Skeleton length: The spatial distance from a feature point along the skeleton to its opposite endpoint; Skeleton orientation: The orientation angle of the skeleton associated with the feature point; Average length of the scaffold: The average length of the scaffold adjacent to the main scaffold containing the feature point. Local curvature: The approximate discrete curvature of a feature point on a contour line.

[0094] The starting point selection module is used to determine the optimal matching starting point for the feature point sequences of two contour lines. Specifically, the starting point selection module calculates the matching cost between all pairs of principal feature points. The matching cost integrates spatial distance, difference in backbone length, difference in backbone orientation, difference in average length of branches, and difference in local curvature. The principal feature point pair with the minimum cost is selected as the serialization starting point. The two feature point sequences are rearranged to align the starting points, resulting in the feature point sequences of the two ore body contour lines. and For the detailed formula, please refer to step 4 above.

[0095] The optimal mapping module employs an improved dynamic programming algorithm to solve for the optimal mapping of feature points along the ore body contour, while satisfying one-to-one mapping, order preservation, and global optimality. This is expressed as: .

[0096] One-to-one mapping requires that the mapping be one-to-one, meaning that each selected feature point can be matched at most once, and the number of matched point pairs is equal. Order preservation requires that the mapping does not disrupt the topological order of the contour lines themselves; that is, if the sequence... If one feature point is ranked before another feature point, then they are in the sequence. Corresponding points must also maintain the same order to avoid overlapping correspondences. Optimality requires that the sum of the matching costs of all matching point pairs reach a global minimum. The matching cost is given by the input matrix; the smaller the value, the more similar the two feature points are in terms of geometry and topology.

[0097] The optimal mapping module uses a dynamic programming algorithm to solve for the optimal mapping of feature points on the ore body contour line. Specifically, it first initializes the boundary states, handling the case where only one sequence contains feature points, and accumulating skip penalties sequentially. Then, it calculates each state in ascending order, selecting the transition path with the minimum total cost from three operations: matching the current point, skipping the current point of the first sequence, and skipping the current point of the second sequence. The matching operation introduces a continuous matching reward coefficient to encourage continuous correspondence; the skipping operation introduces an increasing penalty for continuous skipping, making the penalty heavier the more consecutive skips, thereby suppressing unreasonable long-segment skips. See step 5 above for details.

[0098] The model building and output module is used to construct a 3D surface based on the optimal mapping results and output the final model. Specifically, the model building and output module includes: Segmented Unit: Based on the matching feature point pairs in the optimal mapping, the two contour lines are decomposed into multiple non-closed curve segments respectively; Triangular mesh construction unit: For each pair of corresponding curve segments, a triangular mesh is constructed using the synchronous forward method, the minimum area method, or the minimum perimeter method; Merging Units: Merge all segmented triangular meshes to obtain a complete triangular mesh surface between adjacent contour lines; Output Unit: Merges the triangular mesh surfaces of all adjacent ore body outline pairs to form a complete 3D model of the ore body and outputs it.

[0099] The present invention also discloses an electronic device for 3D modeling of ore bodies that takes into account the matching of contour geometric features, comprising at least one processor, a memory communicatively connected to the at least one processor, a user interface, and a network interface.

[0100] The user interface is used to receive ore body contour data or modeling parameters input by the user; the network interface is used to obtain ore body contour data from an external database or output a three-dimensional ore body model; the memory stores a computer program that can be executed by the at least one processor, and when the computer program is executed by the processor, it causes the processor to execute a ore body three-dimensional modeling method based on optimal mapping of contour skeleton as described above.

[0101] In summary, this invention, by integrating the internal and external skeletons, comprehensively represents the morphological features of the contour lines, overcoming the shortcomings of traditional methods that only focus on the internal skeleton while ignoring external expansion morphological information. Based on this, through hierarchical skeleton processing, the main and secondary skeletons are identified from the complex skeleton network, guiding the stable extraction of key feature points. The improved dynamic programming algorithm introduces a continuous matching reward mechanism during feature point matching, encouraging the formation of coherent correspondences and avoiding discontinuous matching results. Simultaneously, the continuous skipping incremental penalty mechanism effectively suppresses unreasonable long-segment skips, preventing the loss of important morphological details. The comprehensive application of these techniques ensures the rationality and geological consistency of point-to-point mapping under complex morphological changes, effectively avoiding geometric errors such as triangular face distortion and self-intersection common in traditional methods, and significantly improving the construction quality of the 3D model.

[0102] The above description is merely an embodiment of the present invention and does not constitute any limitation on the technical scope of the present invention. Therefore, any minor modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A 3D modeling method for ore bodies based on optimal mapping of contour skeletons, wherein a set of approximately parallel ore body contour lines are input, each ore body contour line being a closed polyline on a spatial plane, and a triangular mesh is constructed between adjacent ore body contour lines through wireframes, thereby obtaining a 3D model of the ore body, characterized in that... The method specifically includes the following steps: Step 1: For the outline of adjacent ore bodies and The internal and external skeletons of each ore body contour line were extracted using the wavefront contraction method. Step 2: Classify the extracted skeleton and identify the main and secondary feature points of the ore body outline, collectively referred to as ore body outline feature points; Step 3: Construct feature vectors for each ore body contour feature point. The feature vectors include the length of the main skeleton, the orientation of the main skeleton, the average length of the supporting skeleton, and the local curvature. Step 4: Calculate the matching cost between the main feature points of two adjacent ore body contour lines. The matching cost integrates spatial distance, difference in main skeleton length, difference in main skeleton orientation, difference in average length of supporting skeleton, and difference in local curvature. Select the main feature point pair with the minimum matching cost as the starting point of the feature point sequence, and rearrange the feature point sequence to obtain the ore body contour line. The feature point sequence is ore body outline The feature point sequence is ; Step 5: Use dynamic programming algorithm to solve for the optimal mapping between the feature point sequences of two adjacent ore body contour lines. The optimal mapping satisfies one-to-one mapping, order preservation and global optimality. Step 6: Based on the optimal mapping, decompose the two adjacent ore body contour lines into multiple non-closed curve segments. Construct a triangular mesh for each pair of corresponding curve segments, merge all triangular meshes to obtain the triangular mesh surface between adjacent contour lines, and merge the triangular mesh surfaces of all adjacent contour line pairs to obtain the complete ore body 3D model.

2. The 3D modeling method for ore bodies based on optimal mapping of contour skeleton according to claim 1, characterized in that, Step 1 is described in detail as follows: For the outline of adjacent ore bodies and ore body outline and Perform direction consistency processing; Then, the contour lines of the ore bodies were extracted using the wavefront contraction method. and Internal skeleton and The inner normal direction of the wavefront contour line contracts uniformly inward, recording the motion trajectory of all vertices during the wavefront motion, and the union of all trajectories constitutes the internal skeleton. Next, the contour lines of the ore bodies were extracted using the wavefront contraction method. and external skeleton and First, calculate the convex hull of the contour. Adjust the wavefront motion direction to the outward normal direction. Expand outward with the convex hull boundary as the termination boundary. Record the motion trajectory of all vertices. Their union constitutes the external skeleton.

3. The ore body 3D modeling method based on optimal mapping of contour skeleton according to claim 1, characterized in that, Step 2 is described in detail below: First, the skeleton lines that intersect with the ore body outline or convex hull boundary are defined as the supporting skeleton lines, and the skeleton lines that do not intersect with the ore body outline or convex hull boundary are defined as the main skeleton lines. The internal skeleton of the ore body outline is regarded as an acyclic connected undirected graph. The longest path of the internal skeleton of the ore body outline is obtained by using a depth-first search or breadth-first search algorithm. This longest path is defined as the main skeleton, and the other paths of the internal skeleton of the ore body outline and the paths of the external skeleton are defined as the secondary skeleton. Then, the branches connected to the endpoints of the main skeleton or secondary skeleton are traversed sequentially, and the longest branch connected to each endpoint is merged into the main skeleton or secondary skeleton. The endpoints of the main skeleton are defined as primary feature points, and the endpoints of the secondary skeleton are defined as secondary feature points. The primary and secondary feature points are collectively referred to as ore body contour feature points.

4. The 3D modeling method for ore bodies based on optimal mapping of contour skeleton according to claim 1, characterized in that, In step 3, for each ore body contour feature point The spatial coordinates of the point Its eigenvectors are calculated as follows: skeleton length Feature points The associated skeleton length; Skeletal position Feature points The orientation of the associated skeleton is defined as a vector. The azimuth angle; Average length of the support frame Feature points The average length of the adjacent branch skeletons of the main skeleton; Local features Feature points Local contextual features along the ore body outline are obtained through discrete curvature approximation; let feature points be... The adjacent points on the outline are respectively and Their coordinates are ( )and( If ), then the local feature is defined as: ,in, , , , .

5. The ore body 3D modeling method based on optimal mapping of contour skeleton according to claim 4, characterized in that, In step 4, the matching cost is calculated as follows: In the formula, Represents the outline of the ore body Main feature points Outline of the ore body Main feature points The matching cost, , , and These represent the weights of the trunk skeleton length, trunk skeleton orientation, average length of the branch skeleton, and local features, respectively.

6. The ore body 3D modeling method based on optimal mapping of contour skeleton according to claim 1, characterized in that, In step 5 One-to-one mapping: π is a one-to-one mapping, that is, for any There exists a unique one. Make And inverse mapping The presence of a feature point indicates that each selected feature point can be matched at most once. Order preservation: if ( i <i' ),but ; indicates if exist Middle row Previously, they were in The corresponding points in the text must also maintain the same order: exist Middle row Before; Optimality: The sum of the matching costs of all matching pairs is minimized globally.

7. The ore body 3D modeling method based on optimal mapping of contour skeleton according to claim 1, characterized in that, The dynamic programming algorithm in step 5 is specifically as follows: (1) State definition main state : forward points and forward Minimum total matching cost for points; Auxiliary status :arrive( , )hour The number of consecutive skipped points; Auxiliary status :arrive( , )hour The number of consecutive skipped points; Auxiliary status :arrive( , The number of point pairs matched when ). (2) Boundary condition initialization For all Initialize the main state Auxiliary state Auxiliary state Auxiliary state ; For the first column: , , ; For the first line: , , ; in, This indicates that the basic penalty is skipped. The mean of the cost matrix, This is an adjustment factor, ranging from 0.1 to 0.5; (3) State transition equation for All other states will be calculated step by step during the state transition process, and the calculation equations are as follows: In the formula, under the matching operation, feature points and Establish a correspondence; from the state ( The total cost increases due to the transfer. ,in, This is the consecutive match reward coefficient; after a match, the consecutive skip count is reset. , ; Number of matching pairs ; jump over Feature points in This indicates that feature points will not be included. If it matches any point, skip it directly; from the state ( The total cost increases due to the transfer. ,in, It continuously skips increasing coefficients; it skips feature points. back, , ; jump over Feature points in This indicates that feature points will not be included. If it matches any point, skip it directly; from the state ( The total cost increases due to the transfer. Skip feature points back, , ; Through the above state transition operations, a feature point mapping relationship that satisfies one-to-one mapping, order preservation, and optimality is finally obtained. .

8. A 3D modeling system for ore bodies based on optimal mapping of contour skeletons, characterized in that, include: The data input module is used to input a set of ore body contour lines that are approximately parallel to each other. Each ore body contour line is a closed polyline on a spatial plane, consisting of a series of vertices arranged in sequence. The skeleton extraction module is used to make the orientation of adjacent contour lines consistent and to extract the internal and external skeletons of each ore body contour line. The skeleton extraction module specifically includes: Direction unification unit: For a pair of adjacent ore body contour lines currently being processed, the vertex sequences of the two are processed to unify the direction to either clockwise or counterclockwise, so that the normals of the two contour lines are consistent. Internal skeleton extraction unit: The wavefront contraction method is adopted, and the inner normal direction of the wavefront contour line is uniformly contracted inward. The motion trajectory of all vertices during the wavefront motion is recorded, and the union of all trajectories constitutes the internal skeleton. External skeleton extraction unit: First, calculate the convex hull of the contour line, and then use the same wavefront contraction principle as the internal skeleton, but adjust the wavefront motion direction to the outward normal direction. Expand outward with the convex hull boundary as the termination boundary, record the motion trajectory of all vertices, and their union constitutes the external skeleton. The skeleton hierarchy module is used to identify dominant feature points from the skeleton network. This skeleton hierarchy module includes: Skeleton classification unit: The skeleton line that intersects with the ore body outline or convex hull boundary is defined as the branch skeleton, and the skeleton line that does not intersect with the ore body outline or convex hull boundary is the main skeleton. Longest path search unit: The internal skeleton of the ore body outline is regarded as an acyclic connected undirected graph. The longest path is extracted as the main skeleton using depth-first search or breadth-first search algorithm. Other paths of the internal skeleton of the ore body outline and the paths of the external skeleton are used as secondary skeletons. Support skeleton merging unit: Traverse the support skeletons connected to the endpoints of the main skeleton or secondary skeleton, and merge the longest support skeleton connected to each endpoint into the main skeleton or secondary skeleton, so that the endpoints of the main skeleton fall on the ore body outline or convex hull boundary. Feature point definition unit: The endpoints of the main skeleton are defined as the main feature points of the ore body outline, and the endpoints of the secondary skeleton are defined as the secondary feature points of the ore body outline. The main and secondary feature points of the ore body outline are collectively referred to as the feature points of the ore body outline. The feature vector construction module is used to construct a four-dimensional feature vector for each feature point, quantifying its geometric and topological properties. The four-dimensional feature vector includes: Skeleton length: The spatial distance from a feature point along the skeleton to its opposite endpoint; Skeleton orientation: The orientation angle of the skeleton associated with the feature point; Average length of the scaffold: The average length of the scaffold adjacent to the main scaffold containing the feature point. Local curvature: an approximation of the discrete curvature of a feature point on a contour line; The starting point selection module is used to determine the optimal matching starting point for the feature point sequences of two contour lines. Specifically, the starting point selection module calculates the matching cost between all pairs of main feature points. The matching cost integrates spatial distance, difference in backbone length, difference in backbone orientation, difference in average length of branches, and difference in local curvature. The main feature point pair with the minimum cost is selected as the serialization starting point. The two feature point sequences are rearranged to align the starting points, resulting in the feature point sequences of the two ore body contour lines. and ; The optimal mapping module employs an improved dynamic programming algorithm to solve for the optimal mapping of feature points along the ore body contour, while satisfying one-to-one mapping, order preservation, and global optimality. This is expressed as: . The model building and output module is used to construct a 3D surface based on the optimal mapping results and output the final model. Specifically, the model building and output module includes: Segmented Unit: Based on the matching feature point pairs in the optimal mapping, the two contour lines are decomposed into multiple non-closed curve segments respectively; Triangular mesh construction unit: For each pair of corresponding curve segments, a triangular mesh is constructed using the synchronous forward method, the minimum area method, or the minimum perimeter method; Merging Units: Merge all segmented triangular meshes to obtain a complete triangular mesh surface between adjacent contour lines; Output Unit: Merges the triangular mesh surfaces of all adjacent ore body outline pairs to form a complete 3D model of the ore body and outputs it.

9. A 3D modeling system for ore bodies based on optimal mapping of contour skeletons according to claim 8, characterized in that, In the process of solving the optimal mapping of feature points of the ore body contour line by the optimal mapping module, the one-to-one mapping property requires that the mapping is one-to-one, that is, each selected feature point can be matched at most once, and the number of matched point pairs is equal; the order preservation property requires that the mapping does not destroy the topological order of the contour line itself; the optimality property requires that the sum of the matching costs of all matched point pairs reaches the global minimum. The optimal mapping module uses a dynamic programming algorithm to solve for the optimal mapping of feature points on the ore body contour line. Specifically, it first initializes the boundary states and handles the case where only one sequence contains feature points, accumulating skip penalties sequentially. Then, it calculates each state in ascending order and selects the transition path with the minimum total cost from three operations: matching the current point, skipping the current point of the first sequence, and skipping the current point of the second sequence. Among these operations, the matching operation introduces a continuous matching reward coefficient to encourage the formation of continuous correspondences; the skipping operation introduces an increasing penalty for continuous skipping, so that the more consecutive skips there are, the heavier the penalty becomes, thereby suppressing unreasonable long-segment skips.

10. An electronic device for 3D modeling of ore bodies that takes into account both contour line geometric feature matching, characterized in that: include At least one processor, The memory that is communicatively connected to the at least one processor The user interface is used to receive ore body contour data or modeling parameters input by the user. A network interface is used to obtain ore body contour data from an external database or output a 3D ore body model. The memory stores a computer program that can be executed by the at least one processor. When the computer program is executed by the processor, it causes the processor to perform a three-dimensional modeling method for ore bodies based on optimal mapping of contour skeleton as described in claims 1-7.

Images