Method for identifying abnormal structure in iron ore exploration process based on three-dimensional laser surveying and mapping

Abstract

The invention belongs to the technical field of anomaly recognition, and particularly relates to a method for recognizing an abnormal structure in an iron ore exploration process based on three-dimensional laser surveying and mapping, which comprises the following steps of: acquiring three-dimensional point cloud data of the earth surface of a mining area by using three-dimensional laser surveying and mapping equipment, constructing a structured point cloud data set, and constructing an abnormal structure in an iron ore exploration process based on the constructed structured point cloud data set; constructing a point cloud space graph structure, and determining an abnormal structure candidate region in the iron ore exploration process according to the constructed point cloud space graph structure; and dividing the abnormal structure candidate region into a plurality of sub-regions by using a sub-region division mechanism, and determining an abnormal structure in the iron ore exploration process through a conjoint analysis mechanism for abnormal data nodes in the sub-regions, according to the method, the abnormal structure region merging technology based on the spatial connectivity and the topological relation is adopted, and spatial continuity analysis processing is performed on the abnormal region, so that the finally determined abnormal structure region has good integrity in space.

Description

Technical Field

[0001] This invention belongs to the technical field of anomaly identification, specifically a method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping. Background Technology

[0002] Currently, the identification of abnormal geological structures in the process of iron ore exploration mainly relies on on-site investigations by geologists, two-dimensional geological profile analysis, and surface morphology interpretation methods based on three-dimensional laser mapping data. By analyzing the spatial characteristics of the surface elevation undulations, fault traces, and abrupt topographic changes in the mining area, the occurrence conditions of underground iron ore and the distribution of abnormal structures can be inferred.

[0003] However, existing technologies for accurately identifying anomalous structures based on 3D laser mapping data are mainly limited by the weak correlation between surface point cloud data and underground geological structures, the reliance on human experience in expressing anomalous structure features, and the difficulty in effectively suppressing noise interference under complex terrain conditions. As a result, it is difficult to balance the accuracy of anomalous structure identification with the efficiency of rapid automated analysis in large-scale mining areas in practical applications, thus affecting the overall reliability of iron ore exploration and its engineering application value. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping. This method solves the problems in existing technologies, such as the weak correlation between surface point cloud data and underground geological structures, the reliance on human experience in expressing anomalous structure features, and the difficulty in effectively suppressing noise interference under complex terrain conditions.

[0005] A method for identifying anomalous structures in iron ore exploration based on 3D laser mapping includes the following steps: determining candidate regions for anomalous structures during iron ore exploration, specifically:

[0006] Three-dimensional point cloud data of the surface of the mining area is acquired using three-dimensional laser mapping equipment and constructed into a structured point cloud dataset. Based on the constructed structured point cloud dataset, a point cloud spatial map structure is constructed. Based on the constructed point cloud spatial map structure, candidate areas of abnormal structures in the iron ore exploration process are determined.

[0007] Furthermore, the identification of anomalous structures in the iron ore exploration process based on candidate anomalous structure regions is as follows:

[0008] The candidate region for anomalous structures is divided into multiple sub-regions using a sub-region partitioning mechanism. For anomalous data nodes within these sub-regions, a joint analysis mechanism is used to identify anomalous structures in the iron ore exploration process. This also includes:

[0009] A subgraph structure is constructed based on the divided sub-regions. Based on the constructed subgraph structure, the construction consistency feature parameters in the subgraph structure are calculated, including the construction direction consistency feature, the construction morphology continuity feature, and the construction structure stability feature. Abnormal constructions are determined based on the calculated construction consistency feature parameters.

[0010] Preferably, the construction of the point cloud spatial graph structure is as follows:

[0011] Point cloud data in a structured point cloud dataset is described in the form of data nodes. Based on the spatial proximity between data nodes and the geometric similarity of the point cloud, the topological connection relationship between data nodes is determined, and a point cloud spatial graph structure is constructed based on the topological connection relationship between each node.

[0012] Preferably, the determination of the topological connection relationship between data nodes is specifically as follows:

[0013] Calculate the spatial Euclidean distance between any two data nodes and the geometric similarity between the point cloud data corresponding to the data nodes. Determine the topological connectivity based on the calculation results, then:

[0014] By setting spatial Euclidean distance threshold and geometric similarity threshold respectively, and determining the topological connectivity based on the set spatial Euclidean distance threshold and geometric similarity threshold, we have:

[0015] If the spatial Euclidean distance between two data nodes is less than the spatial Euclidean distance threshold, and the geometric similarity between the point cloud data corresponding to the two data nodes is less than the geometric similarity threshold, then it means that a topological connection relationship can be determined between the two data nodes, and there is an edge connection relationship between the two data nodes. Conversely, it means that a topological connection relationship cannot be determined between the two data nodes, and there is no edge connection relationship between the two data nodes.

[0016] Preferably, the determination of candidate areas for anomalous structures during iron ore exploration is specifically as follows:

[0017] For each data node in the point cloud spatial graph structure, the structural feature parameters of each data node are calculated, including node elevation gradient features, node local curvature change features, and node density abrupt change features. The calculated structural feature parameters are then used to construct the structural feature vector of the corresponding data node.

[0018] Based on the constructed structural feature vectors, candidate regions for anomalous structures in the iron ore exploration process are determined, and then:

[0019] Extract point cloud data corresponding to normal constructions from the historical operation database and construct normal construction structure feature vectors; based on the constructed normal construction structure feature vectors, filter out abnormal data nodes from the structure feature vectors of data nodes in the point cloud spatial graph structure; determine abnormal construction candidate regions based on the filtered abnormal data nodes.

[0020] Preferably, the process of filtering out abnormal data nodes is as follows:

[0021] For the constructed normal structural feature vectors, for any data node in the point cloud spatial graph structure, calculate the cosine similarity between the two; set a cosine similarity threshold. Based on the set cosine similarity threshold, abnormal data nodes in the point cloud spatial graph structure are determined, then:

[0022] If the cosine similarity is calculated The comparison with the set cosine similarity threshold satisfies the formula. If the expression is true, it indicates that the current data node is an abnormal data node in the point cloud spatial graph structure; otherwise, it indicates that the current data node is a normal data node in the point cloud spatial graph structure.

[0023] Preferably, the step of determining the candidate region for anomaly construction based on the filtered abnormal data nodes is as follows:

[0024] Based on the identified abnormal data nodes, an abnormal data node is randomly selected from all abnormal data nodes and used as the starting data node for constructing the candidate region of anomalies. According to the determined starting data node, the spatial distance between it and each of the remaining abnormal data nodes is calculated, and the abnormal data node with the closest spatial distance is selected as the next data node after the starting data node. When the second data node of the candidate region of anomalies is determined, the spatial distance between it and each of the remaining abnormal data nodes is calculated again, and the abnormal data node with the closest spatial distance is selected as the third data node of the candidate region of anomalies. This process continues until all abnormal data nodes are used as data in the candidate region of anomalies. Each data node is connected sequentially according to its order, and the region formed by all data nodes is the candidate region of anomalies.

[0025] Preferably, the sub-region partitioning mechanism divides the candidate region for anomaly construction into multiple sub-regions, as follows:

[0026] Based on the identified abnormal data nodes, randomly select one abnormal data node from the list and calculate the spatial distance between the selected abnormal data node and the remaining abnormal data nodes. Sort the calculated spatial distances in ascending order. At the same time, determine a circular area by taking the selected abnormal data node, the nearest abnormal data node in terms of spatial distance, and the next nearest abnormal data node in terms of spatial distance. The determined circular area is the abnormal structural area in the iron ore exploration process.

[0027] Preferably, the construction consistency feature parameters in the computational subgraph structure are as follows:

[0028] The construction direction consistency feature is applied to each abnormal data node in the subgraph structure. Based on its neighboring abnormal data nodes, the local principal direction vector is calculated as follows:

[0029] Calculating the local principal direction vector involves performing principal component analysis on the neighborhood point cloud data of the abnormal data node, extracting the feature vector corresponding to the largest eigenvalue, and using it as the local construction principal direction vector of the abnormal data node. Based on the local construction principal direction vectors of all abnormal data nodes in the subgraph structure, the consistency feature of the construction direction is calculated.

[0030] The structural morphological continuity feature is the degree of continuity in the spatial distribution of outlier data nodes within a divided sub-region, specifically:

[0031] For abnormal data nodes in the subgraph structure, the shortest path connected component is calculated, and the morphological continuity feature is constructed based on the ratio of the number of nodes in the connected component to the total number of nodes in the candidate region.

[0032] Constructing structural stability features involves calculating the degree of dispersion of structural feature vectors based on the structural feature vectors of abnormal data nodes in the subgraph structure, and using the calculation results as the structural stability features.

[0033] Preferably, the determination of anomalous constructions based on computational construction consistency feature parameters is as follows:

[0034] Based on the characteristics of consistent construction direction, continuous construction morphology, and stable construction structure, anomaly construction discrimination feature vectors are constructed. The anomalous construction is determined based on the constructed anomalous construction discrimination feature vectors by comparing the discrimination feature vectors corresponding to known anomalous constructions in the historical operation database with the constructed anomalous construction discrimination feature vectors, and the anomalous construction is determined based on the comparison results.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. This invention adopts an anomaly structure identification technology based on three-dimensional laser mapping point cloud to construct a spatial map structure. It organizes the surface point cloud data of the mining area in the form of data nodes and their topological connection relationships, realizing a structured expression of surface spatial morphological changes. It effectively solves the technical problem that point cloud data in the prior art only serves as discrete geometric information and is difficult to reflect the spatial correlation of geological structures, thus improving the spatial expression capability of anomaly structure identification.

[0037] 2. This invention employs an anomaly data node screening technique based on the comparison of node structural feature vectors with normal structural features. At the point cloud spatial graph structure level, it performs multi-dimensional structural feature analysis on each data node, including elevation gradient, local curvature changes, and density mutations. Combined with normal structural features, it performs similarity discrimination, thereby achieving automatic identification of anomaly nodes. This overcomes the problem of excessive reliance on human experience for anomaly structure interpretation in existing technologies, and improves the objectivity and automation of anomaly structure identification.

[0038] 3. This invention adopts a two-stage identification technology of "abnormal structure candidate region - abnormal structure region". First, abnormal structure candidate regions are determined at the node level, and then normal structures are eliminated and abnormal structures are finely screened at the region level. This achieves a gradual convergence from a large range of candidate regions to a small range of abnormal structure regions, effectively solving the problems of excessively large abnormal identification range and high misjudgment rate in the prior art, and improving the accuracy of abnormal structure localization.

[0039] 4. This invention adopts a normal structure removal technique based on regional structural features to model the regional structural features of local sub-regions within the candidate region of abnormal structures and compares and analyzes them with the features of historical normal structures. This achieves effective removal of normal structures within the candidate region, avoids interference from normal terrain undulations or noisy point clouds on the judgment results of abnormal structures, and significantly improves the reliability of abnormal structure identification results.

[0040] 5. This invention adopts anomaly structure region merging technology based on spatial connectivity and topological relationships to perform spatial continuity analysis and merging processing on anomalous local sub-regions after removing normal structures. This makes the finally determined anomalous structure regions have good spatial integrity and engineering interpretability, effectively solving the problem that the anomaly identification results in the prior art are scattered and difficult to directly guide iron ore exploration deployment, and improving the engineering application value of anomalous structure results. Attached Figure Description

[0041] Figure 1 This is a schematic diagram of the overall method steps of the present invention for identifying abnormal structures in iron ore exploration based on three-dimensional laser mapping. Detailed Implementation

[0042] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0043] Example 1

[0044] Reference Figure 1 As an embodiment of the present invention, a method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping is provided, comprising the following steps:

[0045] S1: Identify candidate areas for anomalous structures during iron ore exploration.

[0046] Specifically, identifying candidate areas for anomalous structures during iron ore exploration involves acquiring 3D point cloud data of the surface of the mining area using 3D laser mapping equipment, and then determining candidate areas for anomalous structures based on the acquired 3D point cloud data. The specific implementation is as follows:

[0047] Acquiring 3D point cloud data of the mining area surface using 3D laser mapping equipment involves using this equipment within the iron ore exploration area to obtain 3D point cloud data of the mining area surface, and then processing the acquired 3D point cloud data to construct a structured point cloud dataset. Specifically:

[0048] Using 3D laser mapping equipment to obtain 3D point cloud data of the mining area surface, we have:

[0049]

[0050] in, Indicates the first Spatial coordinates of a 3D point cloud data This represents the total amount of point cloud data acquired, which is set by the implementers based on the actual application scenario. This represents the acquired set of 3D point cloud data;

[0051] After processing the acquired 3D point cloud data, we have:

[0052] For the acquired 3D point cloud data set Noise removal, coordinate unification, and density equalization are performed to construct a structured point cloud dataset. Noise removal, coordinate unification, and density equalization are all well-known techniques in the art and will not be described in detail in this embodiment.

[0053] Based on the constructed structured point cloud dataset, candidate regions for anomalous structures in the iron ore exploration process are identified. The specific implementation is as follows:

[0054] Based on the constructed structured point cloud dataset, a point cloud spatial graph structure is built, specifically as follows:

[0055] It uses point cloud data from a structured point cloud dataset, describing it in the form of data nodes. Based on the spatial proximity and geometric similarity between the data nodes, the topological connections between them are determined. A point cloud spatial graph structure is then constructed based on these topological connections. Specifically:

[0056] Based on a structured point cloud dataset, a point cloud spatial graph structure is constructed, then:

[0057]

[0058] in, This represents the structure of the constructed point cloud. This represents the set of data nodes in a point cloud spatial graph structure, consisting of point cloud data from a structured point cloud dataset. The edge connections representing the spatial graph structure of a point cloud are determined by the topological connections between data nodes in the point cloud spatial graph structure, specifically:

[0059] For data nodes in a point cloud spatial graph structure, calculate the spatial Euclidean distance between any two data nodes and the geometric similarity between the corresponding point cloud data. Based on the calculation results, determine the topological connectivity. Then, we have...

[0060] The implementation personnel set spatial Euclidean distance thresholds and geometric similarity thresholds according to the actual application scenario. Based on these thresholds, the topological connectivity is determined, resulting in:

[0061] If the spatial Euclidean distance between two data nodes is less than the spatial Euclidean distance threshold, and the geometric similarity between the point cloud data corresponding to the two data nodes is less than the geometric similarity threshold, then it means that a topological connection relationship can be determined between the two data nodes, and there is an edge connection relationship between the two data nodes. Conversely, it means that a topological connection relationship cannot be determined between the two data nodes, and there is no edge connection relationship between the two data nodes.

[0062] Based on the constructed point cloud spatial map structure, candidate regions for anomalous structures in the iron ore exploration process are identified, specifically:

[0063] For each data node in the point cloud spatial graph structure, structural feature parameters are calculated, including node elevation gradient features, node local curvature change features, and node density abrupt change features. These calculated structural feature parameters are then used to construct the corresponding structural feature vector for each data node. Specifically:

[0064] For data nodes in a point cloud spatial graph structure The node elevation gradient characteristics, node local curvature variation characteristics, and node density abrupt change characteristics of the data nodes are calculated respectively, specifically as follows:

[0065] The elevation gradient characteristics of the nodes are as follows:

[0066]

[0067] The characteristics of local curvature variation at nodes are as follows:

[0068]

[0069] The abrupt change characteristics of node density are as follows:

[0070]

[0071] Based on the calculated node elevation gradient features, node local curvature variation features, and node density abrupt change features, a structural feature vector is constructed, resulting in:

[0072]

[0073] in, Represents data nodes The set of neighboring nodes, Represents data nodes The eigenvalues ​​of the next neighborhood covariance matrix Represents data nodes Local spatial volume, Represents data nodes The nodal elevation gradient characteristics, Represents data nodes The characteristics of local curvature changes, Represents data nodes The node density abrupt change characteristics, Represents data nodes The corresponding structural feature vector.

[0074] Based on the structural feature vectors of data nodes in the point cloud spatial graph structure, candidate regions for anomalous structures in the iron ore exploration process are determined as follows:

[0075] Extract point cloud data corresponding to normal construction from the historical operation database, and calculate the corresponding structural feature parameters based on the extracted point cloud data corresponding to normal construction. Construct the normal construction structural feature vector from the calculated structural feature parameters.

[0076] Based on the constructed normal structural feature vectors, abnormal data nodes are selected from the structural feature vectors of data nodes in the point cloud spatial graph structure, as follows:

[0077] Based on the constructed normal structural feature vector, calculate the cosine similarity between the structural feature vector of each data node in the point cloud spatial graph structure and the normal structural feature vector. Then, determine the abnormal data nodes based on the calculated cosine similarity. Specifically:

[0078] For the constructed normal structural feature vector For any data node in the point cloud spatial graph structure Corresponding structural feature vectors Calculating the cosine similarity between the two, we have:

[0079]

[0080] in, This represents the feature vector of a normal construction structure. Represents data nodes The corresponding structural feature vectors, The calculated cosine similarity is used to identify outlier data nodes in the point cloud spatial graph structure, specifically:

[0081] Set cosine similarity threshold Based on the set cosine similarity threshold, abnormal data nodes in the point cloud spatial graph structure are determined, then:

[0082] If the calculated cosine similarity, when compared with the set cosine similarity threshold, satisfies the formula... If the expression is true, it indicates that the current data node is an abnormal data node in the point cloud spatial graph structure; otherwise, it indicates that the current data node is a normal data node in the point cloud spatial graph structure.

[0083] Based on the selected abnormal data nodes, candidate regions for anomaly construction are determined, specifically as follows:

[0084] Based on the identified abnormal data nodes, an abnormal data node is randomly selected from all abnormal data nodes and used as the starting data node for constructing the candidate region of anomalies. According to the determined starting data node, the spatial distance between it and each of the remaining abnormal data nodes is calculated, and the abnormal data node with the closest spatial distance is selected as the next data node after the starting data node. When the second data node of the candidate region of anomalies is determined, the spatial distance between it and each of the remaining abnormal data nodes is calculated again, and the abnormal data node with the closest spatial distance is selected as the third data node of the candidate region of anomalies. This process continues until all abnormal data nodes are used as data in the candidate region of anomalies. Each data node is connected sequentially according to its order, and the region formed by all data nodes is the candidate region of anomalies.

[0085] S2: Identify anomalous structures in the iron ore exploration process based on candidate regions of anomalous structures.

[0086] Specifically, identifying anomalous structures in iron ore exploration based on candidate anomalous structure regions involves dividing the identified candidate regions into multiple sub-regions using a sub-region division mechanism. For each sub-region, a joint analysis mechanism is then used to determine the anomalous structures in the iron ore exploration process. The specific implementation is as follows:

[0087] The anomaly construction candidate region is divided into multiple sub-regions using a sub-region partitioning mechanism. This sub-region partitioning is performed using a three-point mechanism, specifically:

[0088] Based on the identified anomalous data nodes, a random anomalous data node is selected from the list. The spatial distance between the selected anomalous data node and the remaining anomalous data nodes is calculated, and the calculated spatial distances are sorted in ascending order. Simultaneously, a circular region is determined using the selected anomalous data node, its nearest spatially distant anomalous data node, and the next nearest spatially distant anomalous data node. This determined circular region represents the anomalous structural region during iron ore exploration. The specific implementation is as follows:

[0089] Construct a set of abnormal data nodes from the identified abnormal data nodes. And select one abnormal data node from the abnormal data nodes. And calculate the remaining abnormal data nodes. The spatial distance between them is:

[0090]

[0091] in, Indicates abnormal data nodes 3D point cloud data, Indicates abnormal data nodes 3D point cloud data, This represents the calculated spatial distance, and the calculated spatial distances are sorted in ascending order to obtain the distance sequence of abnormal data nodes;

[0092] Select the abnormal data nodes from the sorted abnormal data nodes. The closest abnormal data node in space and the second closest abnormal data node in spatial distance. abnormal data nodes , and Using three points on the circumference, a circular region is determined. This circular region is then a sub-region of the candidate region for anomaly construction. This process of dividing the region into sub-regions continues until all the abnormal data nodes in the set of abnormal data nodes have been divided into sub-regions. This indicates that the sub-region division of the candidate region for anomaly construction is complete.

[0093] Furthermore, the joint analysis mechanism is used to identify anomalous structures in the iron ore exploration process. This involves targeting anomalous data nodes within a sub-region and using the joint analysis mechanism to determine these anomalous structures. The specific implementation is as follows:

[0094] The subgraph structure is constructed based on the partitioned sub-regions. This process determines the subgraph structure based on the outlier data nodes within each sub-region and the edge connections between these outlier data nodes. Specifically:

[0095] For each sub-region, the anomalous data nodes within that sub-region are identified as the set of anomalous data nodes for anomaly construction identification, and satisfy the formula... And by extracting the edge connections between abnormal data nodes and constructing a subgraph structure, we have:

[0096]

[0097] in, This represents the structure of the constructed subgraph. This represents the set of anomalous data nodes identified through anomaly construction. The edge connection relationships between abnormal data nodes are represented by the edge connection inheritance from the point cloud spatial graph structure.

[0098] Based on the constructed subgraph structure, abnormal constructions of the subgraph structure are identified through a joint analysis mechanism. This involves calculating the construction consistency characteristic parameters in the subgraph structure, including construction direction consistency characteristics, construction morphological continuity characteristics, and construction structural stability characteristics. Abnormal constructions are then identified based on these calculated construction consistency characteristic parameters, as detailed below:

[0099] The construction direction consistency feature is applied to each abnormal data node in the subgraph structure. Based on its neighboring abnormal data nodes, the local principal direction vector is calculated as follows:

[0100] Calculating the local principal direction vector involves performing principal component analysis on the neighborhood point cloud data of the abnormal data node, extracting the eigenvector corresponding to the largest eigenvalue, and using it as the local construction principal direction vector for the abnormal data node.

[0101] Based on the local construction principal direction vectors of all abnormal data nodes in the subgraph structure, the consistency feature of the construction direction is calculated, and then:

[0102]

[0103] in, This indicates the consistency characteristic of the construction direction in computation. , These represent abnormal data nodes. The local construction of the main direction vector of adjacent abnormal data nodes.

[0104] The structural morphological continuity feature is the degree of continuity in the spatial distribution of outlier data nodes within a divided sub-region, specifically:

[0105] For abnormal data nodes in the subgraph structure, calculate the shortest path connected components, and construct morphological continuity features based on the ratio of the number of nodes in the connected components to the total number of nodes in the candidate regions. Then:

[0106]

[0107] in, This indicates the continuity of the calculated structural form. This represents the set of anomalous data nodes within the largest connected component of the subgraph structure.

[0108] Constructing structural stability features involves calculating the dispersion of structural feature vectors from outlier data nodes in the subgraph structure, and using the results as the structural stability features. Specifically:

[0109]

[0110] in, This represents the stability characteristics of the constructed structure as calculated. Represents abnormal data nodes in the subgraph structure. The corresponding structural feature vectors, This represents the mean of the structural feature vectors of all outlier data nodes in the subgraph structure.

[0111] The determination of anomalous structures based on computational construction consistency feature parameters involves constructing anomaly construction discrimination feature vectors based on construction direction consistency features, construction morphological continuity features, and construction structural stability features. Anomalous structures are then determined based on these constructed feature vectors. Specifically:

[0112] Based on the characteristics of consistent construction direction, continuous construction morphology, and stable construction structure, anomaly construction discrimination feature vectors are constructed, resulting in:

[0113]

[0114] in, This represents the constructed anomaly discriminant feature vector. This represents the consistency feature of construction direction in the anomaly construction discriminant feature vector. This represents the continuity feature of the construction morphology in the anomaly construction discriminant feature vector. This represents the structural stability feature in the anomaly construction discrimination feature vector.

[0115] Determining anomaly constructions based on the constructed anomaly construction discrimination feature vector involves comparing the discrimination feature vectors corresponding to known anomaly constructions in the historical operational database with the constructed anomaly construction discrimination feature vector, and determining the anomaly construction based on the comparison result. Specifically:

[0116] The discriminant feature vectors corresponding to known anomaly constructs are extracted from the historical operational database. These extracted discriminant feature vectors are then compared with the constructed anomaly construct discriminant feature vectors. The anomaly construct is determined based on the comparison results. Specifically:

[0117] Let the discriminative feature vector corresponding to the known anomalies extracted from the historical operation database be defined as follows: And by comparing the consistency with the constructed anomaly discriminant feature vector, we have:

[0118]

[0119] in, , , The feature weight coefficients are set by the implementers based on the actual application scenario and must satisfy the formula. , This represents the consistency comparison result between the discriminative feature vector corresponding to the known anomaly construction and the constructed anomaly construction discriminative feature vector, used to determine the anomaly construction, specifically:

[0120] Set the threshold for determining the consistency of abnormal constructions. Based on the set abnormal construction consistency judgment threshold, abnormal constructions are determined, then:

[0121] If the calculated consistency comparison result satisfies the set anomaly construction consistency judgment threshold, then... If the value is 0, it means that the structure corresponding to the abnormal data node in the current subgraph structure is an abnormal structure in the iron ore exploration process; otherwise, it means that the structure corresponding to the abnormal data node in the current subgraph structure is a normal structure in the iron ore exploration process.

[0122] It should be noted that the anomalous structures in each sub-region are determined sequentially. After the anomalous structures in all sub-regions are determined, all the determined anomalous structures are combined. The combined result is the anomalous structures in the iron ore exploration process.

[0123] To further verify the beneficial effects of the present invention, a practical application example of the technical solution of the present invention is provided as follows:

[0124] Experimental area: A known iron ore exploration area (approximately 3.5 km²), which includes multiple verified fault structures, fold structures, and non-abnormal topographic relief areas;

[0125] Data acquisition method: Three-dimensional laser mapping equipment was used to acquire point cloud data of the mining area surface; the point cloud density was approximately 8–12 pts / m²; the point cloud data was processed by unifying coordinates, removing noise and equalizing density before being used as experimental input.

[0126] Benchmark: The "real anomalous structural areas" marked by the geological exploration report and drilling results are used as the true value reference for determining accuracy;

[0127] Execution steps: Construct a point cloud spatial graph structure, filter abnormal data nodes based on node structure feature vectors, determine candidate regions for abnormal constructions, refine candidate regions through a three-point circle sub-region division mechanism, and eliminate normal constructions based on construction direction consistency, morphological continuity, and structural stability.

[0128] Comparison of technical solutions: The landform interpretation method based on 3D laser mapping point cloud only uses two-dimensional / local geometric features such as elevation undulation and slope change, and judges anomalies by regular thresholds;

[0129] Data comparison results:

[0130] Comparison indicators Comparison of technical solutions Technical solution of the present invention Anomaly recognition accuracy 71.3% 89.6% Abnormal structure false detection rate 18.7% 6.2% False positive rate 22.5% 8.1% Candidate anomaly area / actual anomaly area ratio 2.6:1 1.2:1 Normal structure rejection rate 63.4% 87.9% Automated recognition rate 40% 85% Time taken per analysis (3.5 km²) 6.2 hours 3.8 hours

[0131] As can be seen from the comparison table above, compared with existing methods for identifying anomalies based on 3D laser mapping point clouds, this invention demonstrates significant advantages in key indicators such as accuracy, false negative rate, and false positive rate. This is mainly due to the invention's construction of a point cloud spatial map structure and the adoption of a two-stage identification mechanism of "determining candidate regions for anomalies—refining the screening of anomalies," which effectively reduces the interference of complex terrain undulations and noisy point clouds on the identification results. Simultaneously, this invention further introduces a regional-level structural feature discrimination and normal structure elimination mechanism within the candidate regions for anomalies, making the finally identified anomaly regions more spatially concentrated and more connected, thereby reducing the fragmentation of anomaly regions. Furthermore, because this invention significantly improves the automation level of anomaly identification, it shortens the overall analysis time while maintaining identification accuracy, verifying the effectiveness and practical value of this invention in iron ore exploration engineering applications.

[0132] Furthermore, if this function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0133] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0134] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0135] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping, characterized in that, Includes the following steps: The candidate areas for anomalous structures in the iron ore exploration process were identified, specifically as follows: Three-dimensional point cloud data of the surface of the mining area is acquired using three-dimensional laser mapping equipment and constructed into a structured point cloud dataset. Based on the constructed structured point cloud dataset, a point cloud spatial map structure is constructed. Based on the constructed point cloud spatial map structure, candidate areas of abnormal structures in the iron ore exploration process are determined. Furthermore, the identification of anomalous structures in the iron ore exploration process based on candidate anomalous structural regions is specifically as follows: The candidate region for anomalous structures is divided into multiple sub-regions using a sub-region partitioning mechanism. For anomalous data nodes within these sub-regions, a joint analysis mechanism is used to identify anomalous structures in the iron ore exploration process. This also includes: A subgraph structure is constructed based on the divided sub-regions. Based on the constructed subgraph structure, the construction consistency feature parameters in the subgraph structure are calculated, including the construction direction consistency feature, the construction morphology continuity feature, and the construction structure stability feature. Abnormal constructions are determined based on the calculated construction consistency feature parameters.

2. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 1, characterized in that: The specific steps for constructing the point cloud spatial graph structure are as follows: Point cloud data in a structured point cloud dataset is described in the form of data nodes. Based on the spatial proximity between data nodes and the geometric similarity of the point cloud, the topological connection relationship between data nodes is determined, and a point cloud spatial graph structure is constructed based on the topological connection relationship between each node.

3. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 2, characterized in that: The determination of the topological connection relationship between data nodes is as follows: Calculate the spatial Euclidean distance between any two data nodes and the geometric similarity between the point cloud data corresponding to the data nodes. Determine the topological connectivity based on the calculation results, then: By setting spatial Euclidean distance threshold and geometric similarity threshold respectively, and determining the topological connectivity based on the set spatial Euclidean distance threshold and geometric similarity threshold, we have: If the spatial Euclidean distance between two data nodes is less than the spatial Euclidean distance threshold, and the geometric similarity between the point cloud data corresponding to the two data nodes is less than the geometric similarity threshold, then it means that a topological connection relationship can be determined between the two data nodes, and there is an edge connection relationship between the two data nodes. Conversely, it means that a topological connection relationship cannot be determined between the two data nodes, and there is no edge connection relationship between the two data nodes.

4. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 3, characterized in that: The process of identifying candidate regions for anomalous structures during iron ore exploration is as follows: For each data node in the point cloud spatial graph structure, the structural feature parameters of each data node are calculated, including node elevation gradient features, node local curvature change features, and node density abrupt change features. The calculated structural feature parameters are then used to construct the structural feature vector of the corresponding data node. Based on the constructed structural feature vectors, candidate regions for anomalous structures in the iron ore exploration process are determined, and then: Extract point cloud data corresponding to normal constructions from the historical operation database and construct normal construction structure feature vectors; based on the constructed normal construction structure feature vectors, filter out abnormal data nodes from the structure feature vectors of data nodes in the point cloud spatial graph structure; determine abnormal construction candidate regions based on the filtered abnormal data nodes.

5. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 4, characterized in that: The process of filtering out abnormal data nodes is as follows: For the constructed normal structural feature vectors, for any data node in the point cloud spatial graph structure, calculate the cosine similarity between the two; set a cosine similarity threshold. Based on the set cosine similarity threshold, abnormal data nodes in the point cloud spatial graph structure are determined, then: If the cosine similarity is calculated The comparison with the set cosine similarity threshold satisfies the formula. If the expression is true, it indicates that the current data node is an abnormal data node in the point cloud spatial graph structure; otherwise, it indicates that the current data node is a normal data node in the point cloud spatial graph structure.

6. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 5, characterized in that: The process of determining candidate regions for anomaly construction based on filtered abnormal data nodes is as follows: Based on the identified abnormal data nodes, an abnormal data node is randomly selected from all abnormal data nodes and used as the starting data node for constructing the candidate region of anomalies. According to the determined starting data node, the spatial distance between it and each of the remaining abnormal data nodes is calculated, and the abnormal data node with the closest spatial distance is selected as the next data node after the starting data node. When the second data node of the candidate region of anomalies is determined, the spatial distance between it and each of the remaining abnormal data nodes is calculated again, and the abnormal data node with the closest spatial distance is selected as the third data node of the candidate region of anomalies. This process continues until all abnormal data nodes are used as data in the candidate region of anomalies. Each data node is connected sequentially according to its order, and the region formed by all data nodes is the candidate region of anomalies.

7. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 6, characterized in that: The sub-region partitioning mechanism divides the candidate region for anomaly construction into multiple sub-regions, as detailed below: Based on the identified abnormal data nodes, randomly select one abnormal data node from the list and calculate the spatial distance between the selected abnormal data node and the remaining abnormal data nodes. Sort the calculated spatial distances in ascending order. At the same time, determine a circular area by taking the selected abnormal data node, the nearest abnormal data node in terms of spatial distance, and the next nearest abnormal data node in terms of spatial distance. The determined circular area is the abnormal structural area in the iron ore exploration process.

8. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 7, characterized in that: The construction consistency feature parameters in the computational subgraph structure are as follows: The construction direction consistency feature is applied to each abnormal data node in the subgraph structure. Based on its neighboring abnormal data nodes, the local principal direction vector is calculated as follows: Calculating the local principal direction vector involves performing principal component analysis on the neighborhood point cloud data of the abnormal data node, extracting the feature vector corresponding to the largest eigenvalue, and using it as the local construction principal direction vector of the abnormal data node. Based on the local construction principal direction vectors of all abnormal data nodes in the subgraph structure, the consistency feature of the construction direction is calculated. The structural morphological continuity feature is the degree of continuity in the spatial distribution of outlier data nodes within a divided sub-region, specifically: For abnormal data nodes in the subgraph structure, the shortest path connected component is calculated, and the morphological continuity feature is constructed based on the ratio of the number of nodes in the connected component to the total number of nodes in the candidate region. Constructing structural stability features involves calculating the degree of dispersion of structural feature vectors based on the structural feature vectors of abnormal data nodes in the subgraph structure, and using the calculation results as the structural stability features.

9. The method for identifying anomalous structures in iron ore exploration based on three-dimensional laser mapping as described in claim 8, characterized in that: The determination of anomalous constructions based on computational consistency feature parameters is as follows: Based on the characteristics of consistent construction direction, continuous construction morphology, and stable construction structure, anomaly construction discrimination feature vectors are constructed. The anomalous construction is determined based on the constructed anomalous construction discrimination feature vectors by comparing the discrimination feature vectors corresponding to known anomalous constructions in the historical operation database with the constructed anomalous construction discrimination feature vectors, and the anomalous construction is determined based on the comparison results.

Images