Mineral resource potential prediction system and method based on data mining

Abstract

The invention relates to the technical field of mineral resource prediction, in particular to a mineral resource potential prediction system and method based on data mining. The method comprises the following steps of collecting original exploration data, fusing the original exploration data into multi-source geological observation data, dividing the multi-source geological observation data into a plurality of regional geological data according to regions, extracting characteristic variables in the regional geological data, and reconstructing geological evolution characteristics based on the characteristic variables and the regional data, according to the method, mineral resource change data of each region is predicted through geological evolution characteristics, resource change characteristics are identified, potential inversion is performed based on the resource change characteristics, and mineral resource potential is mined in combination with an inversion result, so that more effective resource development and management are realized. According to the method, comprehensive integration of geological information is realized, the prediction precision can be gradually improved in continuous data updating, and the efficiency of mineral resource evaluation is integrally improved.

Description

Technical Field

[0001] This invention relates to the field of mineral resource prediction technology, and in particular to a mineral resource potential prediction system and method based on data mining. Background Technology

[0002] Traditional mineral resource exploration methods often rely on experience and manual analysis. Faced with complex geological conditions and rapidly increasing data volumes, these methods are inadequate, resulting in low data processing efficiency and insufficient accuracy. In particular, the lack of effective fusion technology in integrating multi-source geological data leads to the omission of important information and biased analysis results. Existing technologies have limited capabilities in data mining and feature extraction, failing to fully explore potential mineral resource information and severely restricting the sustainable development of resources. In the process of extracting feature variables and reconstructing geological evolution characteristics, existing methods often fail to fully consider the diversity and complexity of geological data, leading to a decline in the accuracy and reliability of prediction models. Furthermore, the correlation analysis between geological evolution characteristics and mineral resource changes usually lacks systematicity, resulting in the inability to effectively identify resource change characteristics and affecting the scientific validity of potential inversion results. Summary of the Invention

[0003] Therefore, it is necessary to provide a data mining-based mineral resource potential prediction system and method to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a data mining-based method for predicting mineral resource potential includes the following steps: Step S1: Collect raw exploration data; merge the raw exploration data into multi-source geological observation data, and divide the multi-source geological observation data into multiple regional geological data according to each region; Step S2: Extract feature variables from the geological data of each region; reconstruct geological evolution characteristics based on the feature variables and regional geological data; Step S3: Predict mineral resource change data for each region based on geological evolution characteristics; identify resource change characteristics in the mineral resource change data for each region; Step S4: Perform potential inversion based on resource change characteristics, and combine the potential inversion results with resource change characteristics to explore mineral resource potential.

[0005] The present invention also provides a data mining-based mineral resource potential prediction system for performing the data mining-based mineral resource potential prediction method described above. The data mining-based mineral resource potential prediction system includes: The data acquisition module is used to collect raw exploration data; it integrates the raw exploration data into multi-source geological observation data, and divides the multi-source geological observation data into multiple regional geological data according to each region. The feature reconstruction module is used to extract feature variables from geological data of various regions; and to reconstruct geological evolution features based on the feature variables and regional geological data. The evolution identification module is used to predict mineral resource change data in various regions based on geological evolution characteristics; and to identify the resource change characteristics of mineral resource change data in various regions. The potential mining module is used to perform potential inversion based on resource change characteristics, and combine the potential inversion results with resource change characteristics to explore mineral resource potential.

[0006] This invention achieves comprehensive integration of geological information by collecting raw exploration data and fusing it into multi-source geological observation data, providing a solid data foundation for subsequent analysis. Regional division enables the effective differentiation of different geological features, facilitating targeted research. The extraction of characteristic variables provides an important basis for the reconstruction of geological evolution characteristics, allowing for in-depth analysis of the changing trends of geological features in various regions and laying the foundation for resource prediction. Mineral resource change data prediction based on geological evolution characteristics can effectively identify the resource change characteristics of each region and reveal the dynamic change law of mineral resources.

[0007] By combining the results of potential inversion with the characteristics of resource changes, potential mineral resources can be extracted from the data, improving the accuracy and reliability of resource assessment. The modular design of the entire system ensures efficient collaboration among functional modules, enhances the system's flexibility and scalability, and enables it to quickly adapt to resource prediction needs under different geological conditions. The application of data mining technology gives the system self-learning capabilities, allowing it to gradually improve prediction accuracy through continuous data updates. Overall, this improves the efficiency of mineral resource assessment, reduces the need for human intervention, and ultimately achieves scientific prediction of mineral resource potential. Attached Figure Description

[0008] Figure 1 This is a flowchart illustrating the steps of a data mining-based method for predicting mineral resource potential. Figure 2 for Figure 1 A flowchart illustrating step S4; Figure 3 Aerial image of the target area; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0009] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0010] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0011] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated items listed.

[0012] To achieve the above objectives, please refer to Figures 1 to 3 A method for predicting mineral resource potential based on data mining includes the following steps: Step S1: Collect raw exploration data; merge the raw exploration data into multi-source geological observation data, and divide the multi-source geological observation data into multiple regional geological data according to each region; Step S2: Extract feature variables from the geological data of each region; reconstruct geological evolution characteristics based on the feature variables and regional geological data; Step S3: Predict mineral resource change data for each region based on geological evolution characteristics; identify resource change characteristics in the mineral resource change data for each region; Step S4: Perform potential inversion based on resource change characteristics, and combine the potential inversion results with resource change characteristics to explore mineral resource potential.

[0013] In this embodiment, please refer to Figure 3The target area was simultaneously acquired at multiple points using exploration equipment. Seismic reflection signals, gravity survey lines, magnetic profiles, and remote sensing images were used as the main sources of raw exploration data. A multi-channel receiving system was used to record geological response signals at different frequency bands. Time synchronization and spatial registration were performed on each type of raw data to eliminate displacement deviations and noise interference generated during the acquisition process. Subsequently, the multi-source data, including seismic, gravity, magnetic, and remote sensing data, were fused using a layered fusion algorithm to obtain multi-source geological observation data. Based on zoning and geological structural boundaries, the multi-source geological observation data was divided into several regional geological data to form a regional geological dataset containing stratigraphic attributes, lithological types, and structural morphology.

[0014] Feature variables, including multidimensional variables such as stratigraphic thickness, strata dip angle, density gradient, and magnetization, are extracted from geological data of various regions. The coupling degree between variables is calculated using a feature reduction algorithm based on principal component decomposition to select representative key feature variable data. The selected feature variables are spatially registered with the original regional geological data, and a regional spatial feature matrix is ​​formed through three-dimensional interpolation. Geological evolution features are then reconstructed based on the regional spatial feature matrix. During the reconstruction process, the sedimentary layer stacking and tectonic evolution process is simulated based on the time series tomography method, thereby generating a geological evolution feature dataset.

[0015] By predicting mineral resource change data in various regions through geological evolution characteristics, the geological evolution characteristics are input into a prediction model based on stratigraphic evolution weights to calculate the mineralization probability distribution of different stratigraphic units, generating mineral resource change data. Then, resource change characteristics are identified from the mineral resource change data. By calculating parameters such as change rate, enrichment gradient, and spatial correlation, resource change characteristic data describing the trend of resource change are extracted, thus forming a resource change characteristic set containing spatial distribution and temporal evolution laws.

[0016] Based on resource change characteristics, potential inversion is performed. First, resource change field data is constructed, and a multi-scale change field is established with change intensity as the main input parameter. Inversion operator model is applied to the change field data to perform inversion calculation and obtain potential inversion results. The potential inversion results are mapped with resource change characteristics in a unified spatial grid. The potential value and feature correlation coefficient of each unit in the grid are calculated to form fused potential feature data. Then, resource potential distribution data is generated through spatial smoothing. Finally, information on high-potential areas is extracted based on the resource potential distribution data to determine the mineral resource potential of the target area.

[0017] Preferably, step S2, reconstructing the geological evolution characteristics based on characteristic variables and regional geological data, includes: The feature variables are arranged by time index, and the time change rate of the feature variables in the regional geological data is calculated to generate time series sedimentary features; Construct a spatial grid structure based on the spatial variables in the feature variables; The differences between the variables at each node are calculated based on the spatial grid structure to construct the spatial morphological field; The spatial morphological field is formed by calculating the difference between variables at each node and then fused with the spatial morphological field. The evolutionary driving force is simulated to generate a geological evolution field. Identifying geological evolution characteristics based on geological evolution fields.

[0018] In this embodiment, the reconstruction of geological evolution features based on feature variables and regional geological data includes: firstly, arranging feature variables by time index, establishing a time series index with a sampling time interval of 10,000 years as the step size, arranging variables such as stratum thickness, sediment grain size, lithological density and porosity in chronological order to form a multidimensional time series matrix, calculating the rate of change of feature variables between adjacent time layers to obtain a time rate of change sequence, and generating time series sedimentary features with the rate of change as the core parameter.

[0019] A spatial grid structure is constructed based on the spatial variables in the characteristic variables. Regional geological data is projected onto a unified three-dimensional spatial coordinate system. The spatial grid is divided with a grid unit side length of 100 meters. The average stratigraphic density, seismic reflection intensity and magnetic susceptibility value are calculated for each grid unit to form a spatial node set containing multiple physical attributes, thereby establishing spatial grid structure data.

[0020] The spatial morphological field is constructed by calculating the difference of variables at each node based on the spatial grid structure. The node gradient vector field is determined by calculating the difference in stratum thickness, density and magnetic susceptibility between adjacent nodes. The spatial morphological field data is constructed by using the node gradient direction and amplitude as characteristic parameters. This spatial morphological field reflects the stratum deformation trend and the undulation state of the geological structure.

[0021] The spatial morphological field formed by the calculated differences of variables at each node is fused with the historical geological evolution morphological field. The spatiotemporal superposition algorithm is used to simulate the influence of tectonic stress and sedimentation rate changes on the morphological field, forming continuous spatiotemporal evolution trajectory data. The process of strata uplift, faulting and subsidence is simulated to generate geological evolution field data.

[0022] Based on the geological evolution field, geological evolution characteristics are identified. By extracting the tectonic boundary lines, fold axis surfaces, fault zone distribution and sedimentary center locations in the geological evolution field, and combining them with time series sedimentary characteristics, stratigraphic evolution stages and geological event nodes are determined, forming a complete geological evolution characteristic dataset.

[0023] Preferably, step S4 includes: Construct resource change field data based on resource change characteristics; Invert calculations are performed on resource change field data to generate potential inversion results; Spatial correlation analysis is performed between the potential inversion results and resource change characteristics to form resource potential distribution data; Information on high-potential areas is extracted from resource potential distribution data to obtain mineral resource potential results.

[0024] In this embodiment, resource change field data is constructed based on resource change characteristics. The resource change characteristics obtained in the previous steps are mapped to a unified geological grid according to spatial coordinate index. Change intensity, enrichment rate and spatial gradient are used as the main parameters to establish change feature vectors for each grid node. Continuous change field distribution data is generated using a multi-scale interpolation algorithm. Change characteristics of different time periods are superimposed and smoothed to form resource change field data that describes the spatiotemporal change law of mineral resources.

[0025] Inversion calculations are performed on resource change field data to generate potential inversion results. The resource change field data is input into an inversion framework based on a spatial operator model. Multi-scale potential responses are calculated using the node change rate as an input variable. Local potential values ​​are calculated and fused within each spatial unit to form preliminary inversion data. Then, the abnormal response regions are optimized and adjusted through error correction and boundary constraints to finally generate potential inversion result data containing potential intensity and spatial distribution information.

[0026] Spatial correlation analysis is performed between the potential inversion results and resource change characteristics to form resource potential distribution data. The potential inversion results are then mapped to a spatial grid of resource change characteristics. The correlation coefficient between the potential value and the change characteristics is calculated for each grid cell. Significantly correlated regional cells are selected, and these locally correlated cells are spatially aggregated and smoothed to generate fused potential characteristic data. Finally, spatial interpolation and boundary continuity processing are used to form resource potential distribution data to characterize the spatial continuous distribution of potential changes within the region.

[0027] Information on high-potential areas is extracted from resource potential distribution data to obtain mineral resource potential results. Potential value threshold intervals are calculated from resource potential distribution data, and high-potential threshold areas are determined using the quantile method. Clustering and partitioning analysis is performed on high-potential areas to identify the location of potential center points. Non-metallic areas are eliminated by combining geological structural boundaries and stratigraphic distribution constraints. Finally, mineral resource potential results containing the range of high-potential areas, center coordinates, and potential levels are output.

[0028] Preferably, the inversion calculation of resource change field data includes: Resource change field data is decomposed into multi-scale components to obtain multi-scale resource change components; Calculate the gradients of multi-scale resource change components and construct a change gradient field; Constructing an inversion operator model based on a changing gradient field; The potential inversion results are obtained by simulating the inversion using the inversion operator model and resource change field data.

[0029] In this embodiment, the inversion calculation of resource change field data includes decomposing the resource change field data into multi-scale components to obtain multi-scale resource change components. By performing wavelet decomposition on the resource change field data in the frequency domain, the change features at different spatial scales are extracted respectively. The decomposition levels L1, L2, and L3 correspond to different geological structural scales, where L1 reflects local anomalous changes and L3 reflects overall tectonic changes. The decomposition results are recorded as multi-scale resource change components to distinguish the change patterns at different geological levels.

[0030] The gradients of multi-scale resource change components are calculated, and a change gradient field is constructed. Spatial gradient values ​​and orientation angles of each scale component are calculated between adjacent grid nodes to form a gradient vector field, which reflects the rate of resource change and directional distribution characteristics. Regions with abrupt gradient changes are marked as potential geological fault zones or enrichment boundary areas, and a continuous change gradient field is formed by interpolation at the grid level.

[0031] An inversion operator model is constructed based on a changing gradient field. The directionality and intensity in the gradient field are used as weight inputs. A spatial inversion operator matrix is ​​established using regularization constraints. Inversion weight coefficients and smoothing coefficients are defined for each grid node to ensure that the inversion results retain both local anomaly information and overall continuity, forming an inversion operator model with spatial adaptive characteristics. The model is also regionalized according to geological type parameters to enable the model to have differentiated response capabilities under different geological backgrounds.

[0032] The inversion operator model and resource change field data are used to simulate and invert the potential to obtain the inversion results. The resource change field data is input into the inversion operator model, and the potential response is calculated node by node in the grid space to obtain local potential response data. The results of each local response are then fused in the spatial dimension to form a preliminary potential field. Then, through error correction and boundary smoothing optimization, abnormal peak areas and transitional abrupt areas are removed. Finally, a potential inversion result with good continuity and spatial consistency is obtained to characterize the distribution of mineral resource enrichment potential in the target area.

[0033] Preferably, the simulation and inversion using the inversion operator model and resource change field data includes: Local potential response is calculated based on resource change field data and inversion operator model to form local potential data; Local potential data are fused within a spatial grid to generate a fused potential field; By correcting the anomalies in the fusion potential field, a corrected potential field is obtained; Potential inversion results are extracted based on the corrected potential field.

[0034] In this embodiment, the simulation and inversion using the inversion operator model and resource change field data includes: calculating the local potential response based on the resource change field data and the inversion operator model to form local potential data; inputting each grid node of the resource change field data into the inversion operator model; calculating the potential response value and directional gradient response for the corresponding node; establishing a local response matrix through the rate of change between adjacent nodes to reflect the potential enrichment trend within the geological unit; and normalizing the response matrix to make the response values ​​at different scales comparable, ultimately forming continuously distributed local potential data.

[0035] Local potential data are fused within a spatial grid to generate a fused potential field. The local potential values ​​of adjacent grid nodes are weighted and superimposed according to spatial distance weight and geological continuity weight. Bilinear interpolation is used to smooth local transition zones to avoid high-frequency oscillations. The fusion of potential data is completed through global spatial traversal, so that the fused potential field can reflect the continuous distribution characteristics of resource-rich areas in space, while maintaining the identifiability of local structural anomalies.

[0036] Anomalies in the fusion potential field are corrected to obtain a calibrated potential field. Anomaly high value regions and abrupt boundary regions are detected in the fusion potential field. Anomaly potential values ​​higher than a set threshold are marked as local anomalies and corrected by the neighborhood mean smoothing method. Gradient constraint correction is applied to the boundary abrupt region to make the potential gradient change continuous and consistent in direction, so as to eliminate error interference caused by measurement noise or data discreteness. Finally, a calibrated potential field with a smooth structure and physical rationality is formed.

[0037] Based on the potential inversion results extracted from the corrected potential field, potential contour lines and high-value center areas are extracted from the corrected potential field in different geological zones. Potential peak areas are identified and potential levels are divided according to peak intensity. Distribution data of high, medium and low potential areas are output. Spatial clustering is used to identify the boundaries of connected areas with high potential values, and finally potential inversion results with clear distribution boundaries and intensity classification characteristics are formed.

[0038] Preferably, calculating the local potential response based on resource change field data and the inversion operator model includes: The resource change field data is mapped to the spatial nodes of the inversion operator model to obtain spatial mapping data; Calculate the weighted response of the spatial mapping data at each spatial node to form node response data; Aggregate the node response data according to local neighborhoods to obtain local response summary data; The local response summary data is normalized to generate local potential data.

[0039] In this embodiment, the calculation of local potential response based on resource change field data and inversion operator model includes mapping resource change field data to spatial nodes of inversion operator model. When mapping resource change field data to spatial nodes of inversion operator model, the spatial coordinates of each node are used as indexes to project the change values ​​of corresponding coordinates in resource change field to model nodes proportionally to form spatial mapping data. During the mapping process, bilinear interpolation method is used to reduce discrete error. The spatial resolution is 0.5 meter cells, and the data format is uniformly floating-point array.

[0040] When calculating the weighted response of spatial mapping data at each spatial node, the product of the node's attribute weight and the change value is used as the response value. The weight comes from the coefficient matrix of the inversion operator model and reflects the influence of the node in the global inversion process. The response intensity is calculated by traversing each node and recorded as node response data.

[0041] When aggregating node response data by local neighborhood, a 3×3 or 5×5 neighborhood window is constructed with the node as the center. The mean, variance, and gradient direction of the node response values ​​within the window are statistically analyzed to form local response summary data. The aggregation process is achieved by traversing the entire spatial grid through a sliding window.

[0042] When normalizing the local response summary data, all local response summary values ​​are linearly stretched to the [0,1] interval according to the minimum and maximum values ​​to eliminate the dimensional differences in response amplitude in different regions and obtain local potential data. The normalized data is used for subsequent potential fusion and inversion accuracy analysis, and the data structure maintains a two-dimensional matrix form.

[0043] Preferably, spatial correlation analysis between the potential inversion results and resource change characteristics includes: The potential inversion results are mapped to a spatial grid of resource change characteristics to obtain second spatial mapping data; Calculate the correlation coefficient of the second spatial mapping data within the grid cells to identify locally correlated data; Based on the fusion potential value and change characteristics of locally correlated data, fusion potential feature data is generated; Spatial smoothing is performed on the fusion potential characteristic data to obtain resource potential distribution data.

[0044] In this embodiment, the spatial correlation analysis between the potential inversion results and resource change characteristics includes mapping the potential inversion results to the spatial grid of resource change characteristics. When mapping the potential inversion results to the spatial grid of resource change characteristics, the grid resolution of the resource change characteristics is used as a reference. The coordinate system in the potential inversion results is aligned with the grid coordinate system. The potential inversion values ​​are allocated to the center position of the grid cells through a bilinear interpolation algorithm to obtain the second spatial mapping data. The mapping accuracy is maintained at an error of less than 0.2 grid cells.

[0045] When calculating the correlation coefficient of the second spatial mapping data within the grid cell, the potential inversion value sequence and the resource change characteristic sequence are used as inputs. The Pearson correlation coefficient within each grid cell is calculated to quantify the degree of linear correlation between potential and change characteristics. Areas with a correlation coefficient exceeding 0.6 are identified as significantly correlated areas, thereby identifying locally correlated data. This process is completed within a 64×64 sliding window to balance local characteristics and overall trends.

[0046] When fusing potential values ​​and change characteristics based on local correlation data, the two are weighted according to their correlation strength, with the weight range being [0.3, 0.7]. The fusion potential characteristic value of each grid cell is synthesized by weighted averaging, generating fusion potential characteristic data. This data reflects the common trend between potential distribution and resource changes in space.

[0047] When performing spatial smoothing on the fusion potential feature data, a Gaussian kernel convolution algorithm is used to reduce the impact of noise. The kernel radius is set to 3 grid units. The convolution result forms continuous resource potential distribution data, which is used for subsequent resource trend identification and potential zoning assessment.

[0048] Preferably, calculating the correlation coefficient of the second spatial mapping data within the grid cells to identify locally correlated data includes: The second spatial mapping data is divided into independent grid cells to obtain grid cell data; Within each grid cell, data corresponding to potential values ​​and resource change characteristics are extracted to form cell-to-cell data. The correlation coefficients of the data to the calculated units are used to obtain the correlation coefficient data of the grid units. Based on the correlation coefficient data of grid cells, significantly related cells are selected to form locally correlated data.

[0049] In this embodiment, calculating the correlation coefficient of the second spatial mapping data within the grid cell to identify locally associated data includes dividing the second spatial mapping data into independent grid cells, dividing the data range into grids with a fixed spatial resolution, taking the grid cell side length as 500 meters, and the division result covering the entire potential inversion result and resource change characteristic data area. After division, grid cell data is obtained, and each grid cell records its geographic coordinate boundary and index number.

[0050] When extracting potential values ​​and corresponding data for resource change characteristics within each grid cell, the corresponding pixels in the region are located based on the grid cell index, and the two types of data are extracted to form a set of one-to-one numerical pairs. Each grid cell contains no less than 50 corresponding sample points, forming cell pair data.

[0051] When calculating the correlation coefficient of data between grid cells, the potential value and the resource change characteristic value are used as input data sequences. The linear correlation coefficient is obtained through statistical methods to obtain grid cell correlation coefficient data. This data is used to describe the coupling strength between the potential distribution change and the resource characteristic change within each grid cell. The correlation coefficient value ranges from -1 to 1, with positive values ​​indicating a positive correlation and negative values ​​indicating a negative correlation.

[0052] When screening significantly correlated units based on the correlation coefficient data of grid cells, a correlation coefficient threshold of 0.6 is used as the criterion. Grid cells with correlation coefficients greater than this threshold are identified as significantly correlated areas, and their indices and corresponding coordinate records are used to form local correlation data. This data is used for subsequent spatial fusion and potential feature enhancement analysis to ensure spatial consistency between potential inversion results and resource change characteristics.

[0053] Preferably, the fusion potential value and change characteristics of locally correlated data include: Align the potential inversion results and resource change characteristics in the local correlated data according to grid cells to obtain aligned data; Calculate the potential weight and change feature weight for the aligned data to generate weighted data; Within each grid cell, the fusion potential value is calculated based on weighted data to form grid cell fusion data; The data from the grid cells are integrated into the regional grid to generate fusion potential feature data.

[0054] In this embodiment, the process of fusing potential values ​​and change characteristics based on local correlation data includes aligning the potential inversion results and resource change characteristics in the local correlation data by grid cell. Based on the index of the significantly correlated grid cells recorded in the local correlation data, the numerical records of the corresponding grid cells in the potential inversion result data and the resource change characteristic data are retrieved respectively. Spatial index matching ensures that the two types of data in the same grid cell are completely corresponding in spatial coordinates. The corresponding potential values ​​and change characteristic values ​​are sorted and stored by grid number to obtain aligned data, which contains bivariate information for each grid cell.

[0055] When calculating the potential weight and change feature weight for aligned data, the variance ratio of the two types of data in the local space is used as the basis for weight allocation. The variance of the potential inversion result and the variance of the change feature are normalized into weight coefficients to form the potential weight and the change feature weight, generating weighted data. If the potential change amplitude in a certain grid is higher than the feature change amplitude, the potential weight increases, and vice versa. This method maintains the balance of contributions of the two types of information during fusion, ensuring that the fusion result reflects both the intensity of resource potential and the trend of change.

[0056] When calculating the fusion potential value based on weighted data within each grid cell, the potential value and the change feature value are multiplied by their corresponding weights and then summed to obtain the cell fusion potential value, forming grid cell fusion data. This data is represented as a set of fusion potential indicators within each grid cell, with the value range limited to between 0 and 1, representing the comprehensive potential level of that location, while retaining the original potential and feature fields.

[0057] When integrating grid cell fusion data into a regional grid, all grid cell fusion data are reassembled into a complete regional grid structure based on the grid cell number and regional boundary index. Edge interpolation smoothing is used during the reconstruction process to eliminate abrupt grid boundary changes, ultimately generating fusion potential feature data. This data continuously covers the study area in space and is used to characterize the comprehensive response relationship between potential inversion results and resource change characteristics, providing a high-precision input basis for subsequent resource potential distribution calculations.

[0058] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0059] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A method for predicting mineral resource potential based on data mining, characterized in that, Includes the following steps: Step S1: Collect raw exploration data; The original exploration data was integrated into multi-source geological observation data, and based on the multi-source geological observation data, it was divided into multiple regional geological data according to each region; Step S2: Extract feature variables from the geological data of each region; Reconstruct geological evolution characteristics based on characteristic variables and regional geological data; Step S3: Predict mineral resource change data for each region based on geological evolution characteristics; identify resource change characteristics of mineral resource change data for each region; Step S4: Perform potential inversion based on resource change characteristics, and combine the potential inversion results with resource change characteristics to explore mineral resource potential.

2. The method for predicting mineral resource potential based on data mining according to claim 1, characterized in that, Step S2, which involves reconstructing geological evolution characteristics based on feature variables and regional geological data, includes: The feature variables are arranged by time index, and the time change rate of the feature variables in the regional geological data is calculated to generate time series sedimentary features; Construct a spatial grid structure based on the spatial variables in the feature variables; The differences between the variables at each node are calculated based on the spatial grid structure to construct the spatial morphological field; The spatial morphological field is formed by calculating the difference between variables at each node and fused with the spatial morphological field. The evolutionary driving force is simulated to generate a geological evolution field. Identifying geological evolution characteristics based on geological evolution fields.

3. The method for predicting mineral resource potential based on data mining according to claim 1, characterized in that, Step S4 includes: Construct resource change field data based on resource change characteristics; Invert calculations are performed on resource change field data to generate potential inversion results; Spatial correlation analysis is performed between the potential inversion results and resource change characteristics to form resource potential distribution data; Information on high-potential areas is extracted from resource potential distribution data to obtain mineral resource potential results.

4. The method for predicting mineral resource potential based on data mining according to claim 3, characterized in that, Inversion calculations of resource change field data include: Resource change field data is decomposed into multi-scale components to obtain multi-scale resource change components; Calculate the gradients of multi-scale resource change components and construct a change gradient field; Constructing an inversion operator model based on a changing gradient field; The potential inversion results are obtained by simulating the inversion using the inversion operator model and resource change field data.

5. The method for predicting mineral resource potential based on data mining according to claim 4, characterized in that, Simulation and inversion using inversion operator models and resource change field data includes: Local potential response is calculated based on resource change field data and inversion operator model to form local potential data; Local potential data are fused within a spatial grid to generate a fused potential field; By correcting the anomalies in the fusion potential field, a corrected potential field is obtained; Potential inversion results are extracted based on the corrected potential field.

6. The method for predicting mineral resource potential based on data mining according to claim 5, characterized in that, The calculation of local potential response based on resource change field data and inversion operator model includes: The resource change field data is mapped to the spatial nodes of the inversion operator model to obtain spatial mapping data; Calculate the weighted response of the spatial mapping data at each spatial node to form node response data; Aggregate the node response data according to local neighborhoods to obtain local response summary data; The local response summary data is normalized to generate local potential data.

7. The method for predicting mineral resource potential based on data mining according to claim 3, characterized in that, Spatial correlation analysis between potential inversion results and resource change characteristics includes: The potential inversion results are mapped to a spatial grid of resource change characteristics to obtain second spatial mapping data; Calculate the correlation coefficient of the second spatial mapping data within the grid cells to identify locally correlated data; Based on the fusion potential value and change characteristics of locally correlated data, fusion potential feature data is generated; Spatial smoothing is performed on the fusion potential characteristic data to obtain resource potential distribution data.

8. The method for predicting mineral resource potential based on data mining according to claim 7, characterized in that, Calculating the correlation coefficients of the second spatial mapping data within the grid cells to identify locally correlated data includes: The second spatial mapping data is divided into independent grid cells to obtain grid cell data; Within each grid cell, data corresponding to potential values ​​and resource change characteristics are extracted to form cell-to-cell data. The correlation coefficients of the data to the calculated units are used to obtain the correlation coefficient data of the grid units. Based on the correlation coefficient data of grid cells, significantly related cells are selected to form locally correlated data.

9. The method for predicting mineral resource potential based on data mining according to claim 7, characterized in that, Based on the potential value and change characteristics of locally correlated data fusion, including: Align the potential inversion results and resource change characteristics in the local correlated data according to grid cells to obtain aligned data; Calculate the potential weight and change feature weight for the aligned data to generate weighted data; Within each grid cell, the fusion potential value is calculated based on weighted data to form grid cell fusion data; The data from the grid cells are integrated into the regional grid to generate fusion potential feature data.

10. A mineral resource potential prediction system based on data mining, characterized in that, For executing the data mining-based mineral resource potential prediction method as described in claim 1, the data mining-based mineral resource potential prediction system comprises: The data acquisition module is used to collect raw exploration data; it integrates the raw exploration data into multi-source geological observation data, and divides the multi-source geological observation data into multiple regional geological data according to each region. The feature reconstruction module is used to extract feature variables from geological data of various regions; and to reconstruct geological evolution features based on the feature variables and regional geological data. The evolution identification module is used to predict mineral resource change data in various regions based on geological evolution characteristics; and to identify the resource change characteristics of mineral resource change data in various regions. The potential mining module is used to perform potential inversion based on resource change characteristics, and combine the potential inversion results with resource change characteristics to explore mineral resource potential.

Images