An Artificial Intelligence-Based Geological Prospecting Data Processing and Analysis System

Abstract

This invention relates to the field of data processing technology, specifically to an artificial intelligence-based geological prospecting data processing and analysis system. The system includes: a geological parameter module, a spatial strain module, a causal analysis module, a resonance response module, and an optimized prospecting module. In this invention, spatiotemporal unified expression is achieved through the acquisition and registration of multidimensional geological parameters, eliminating observational biases and improving data comparability. Noise interference is reduced and key feature expression is enhanced through feature dimensionality reduction and ore-forming parameter extraction. Parameter gradient vectorization quantifies strain distribution and reveals tension concentration areas. Directional offset and temporal correlation analysis reveal causal relationships of disturbances and avoid misjudgments. Phase synchronization and amplitude difference analysis of time-series data reflect the energy resonance response of the geological system, forming a hierarchical logic of data fusion, spatial analysis, causal reasoning, and dynamic response, thereby improving the accuracy of geological information analysis and the reliability of prospecting prediction.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an artificial intelligence-based geological prospecting data processing and analysis system. Background Technology

[0002] The field of data processing technology encompasses the entire process of collecting, organizing, transforming, and computing various types of raw data. Its core content involves using computer systems to perform structured management and logical analysis of data, achieving standardized storage, retrieval, and inference support. It typically covers key stages such as data preprocessing, pattern recognition, feature extraction, and result inference, and has a wide range of applications, including industrial control, resource exploration, environmental monitoring, and intelligent decision-making. By integrating and processing data from different sources, data processing technology provides fundamental support for subsequent knowledge mining and model building, and is a crucial supporting direction for achieving intelligent analysis in information systems.

[0003] Among them, the AI-based geological prospecting data processing and analysis system refers to a technical solution that combines artificial intelligence algorithms with geological exploration data processing workflows. For multi-dimensional data such as geophysical, geochemical, and geological structural data generated during geological prospecting, machine learning models are used for data classification and feature recognition, and neural networks are used for geological anomaly zone delineation and mineralization information extraction. Simultaneously, data fusion algorithms are used to complete integrated modeling of multi-source data, and training samples are used for adaptive updating of model parameters, thereby achieving automated analysis and result generation of geological information.

[0004] Current technologies in geological data processing rely on traditional rule-based analysis frameworks, with data fusion often remaining at a static level and lacking the ability to identify the interplay between temporal and spatial differences. Multi-source data often suffers from structural biases due to inconsistent sampling conditions and scales, resulting in the weakening or distortion of geological features in calculations. Geological anomaly identification typically relies on manually set thresholds, failing to reflect the nonlinear and coupled characteristics of geological systems. Causal relationship analysis often relies on simple correlation or regression models, failing to reveal the dynamic interactions between different parameters, leading to a one-sided explanation of mineralization genesis. For disturbance responses in geological processes, traditional methods lack simultaneous analysis at the phase and amplitude levels, easily overlooking the driving effect of resonance characteristics on mineralization. These shortcomings introduce uncertainty into the spatial location and causal reasoning of data analysis results, reducing the accuracy of mineral prediction and increasing exploration costs and the risk of misjudgment. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides an artificial intelligence-based geological prospecting data processing and analysis system, the technical solution of which is as follows:

[0006] The geological parameter module acquires rock density gradient, geothermal gradient, fracture elongation, pore fluid pressure difference, and magnetic field changes and performs spatiotemporal registration. It then performs feature dimensionality reduction on the registered data and extracts the main mineralization parameters, generating a geological parameter set and transferring it to the spatial strain module.

[0007] The spatial strain module performs spatial distribution analysis based on the geological parameter set, calculates and quantifies the gradient difference of geological parameters between adjacent sampling points, analyzes the spatial strain intensity, and if it exceeds the spatial strain threshold, it extracts spatial tension points, generates tension distribution results, and transmits them to the causal analysis module.

[0008] The causal analysis module performs causal chain analysis based on the tension distribution results and the geological parameter set, calculates the directional offset difference and temporal correlation coefficient between multiple geological parameters, filters out areas with reverse and abnormal correlation as disturbance areas and adjusts the causal direction, generates disturbance causal chain results and transmits them to the resonance response module.

[0009] The resonance response module acquires time-series data of multiple geological parameters within the disturbance area, identifies the triggering event points, calculates the phase synchronization rate and resonance coefficient by combining the disturbance causal chain results, evaluates the influence intensity of multiple geological parameters, and generates preliminary mineral exploration results.

[0010] The mineral exploration module is optimized by combining the preliminary mineral exploration results with the perturbation causal chain results, smoothing the distribution of resonance coefficients, eliminating isolated outliers through iterative updates, calculating mineralization factor scores, and generating optimized mineral exploration analysis results.

[0011] As a further aspect of the present invention, the geological parameter set includes rock density gradient, geothermal gradient, and fracture elongation rate; the tension distribution results include spatial strain intensity distribution and spatial tension point information; the disturbance causal chain results include directional offset difference, temporal correlation coefficient distribution, and disturbance region; the preliminary prospecting results include excitation event point set, phase synchronization rate distribution, and resonance coefficient; and the optimized prospecting analysis results include smoothed resonance coefficient, ore-forming factor score, and elimination of isolated anomalies.

[0012] As a further aspect of the present invention, the geological parameter module includes:

[0013] The geological data submodule acquires raw observation data on rock density gradient, geothermal gradient, fracture elongation, pore fluid pressure difference, and magnetic field changes; calculates the variation amplitude and interval average of multiple parameters within the same region; and performs stratigraphic calibration on the parameter sequence to generate the original geological parameter set.

[0014] The spatial registration submodule, based on the original geological parameter set, calls time series and spatial coordinate information to register multiple parameters, calculates the spatial offset and time delay difference of parameters between multiple observation points, and performs coordinate alignment and interpolation correction on multi-source data to generate a spatiotemporal registration parameter set;

[0015] The parameter extraction submodule performs feature dimensionality reduction on rock density gradient, geothermal gradient, fracture elongation rate, pore fluid pressure difference and magnetic field changes based on the spatiotemporal registration parameter set, calculates the contribution rate of multiple parameters in the principal components of mineralization, extracts key mineralization feature sequences, and generates a geological parameter set.

[0016] As a further aspect of the present invention, the space strain module includes:

[0017] The gradient difference calculation submodule performs spatial distribution analysis based on the geological parameter set, extracts the coordinate positions and parameter value sequences of adjacent sampling points, calculates the difference values ​​of geological parameters between multiple points, and standardizes the difference results according to the spatial distance weight to generate a spatial parameter gradient difference set.

[0018] The intensity analysis submodule calls the spatial parameter gradient difference set, calculates the direction vector amplitude of the geological parameter gradient in each region, analyzes the strain intensity value in the region based on the amplitude variation interval, and combines the mean variation of the intensity in adjacent regions for analysis to generate spatial strain intensity distribution results.

[0019] The tension point extraction submodule identifies regions exceeding the spatial strain threshold based on the spatial strain intensity distribution results, extracts the center coordinates and corresponding intensity indices as a set of tension points, calculates the spatial distance and direction distribution of adjacent tension points, and generates spatial tension distribution results.

[0020] As a further aspect of the present invention, the spatial strain threshold is set by obtaining the geological parameter change sequence in multiple typical areas, calculating the spatial parameter gradient difference and the corresponding strain intensity value, and based on the frequency distribution curve of the strain intensity.

[0021] As a further aspect of the present invention, the causal analysis module includes:

[0022] The offset difference calculation submodule, based on the tension distribution result and the spatial coordinate sequence of the geological parameter set, detects the direction vector of adjacent geological parameters and calculates the change in the angle between multiple parameters. Combined with the direction offset reference value, it judges the degree of direction difference between multiple regions and generates a direction offset difference sequence.

[0023] The time series evaluation submodule calls the time series data of the directional offset difference sequence and the geological parameter set, extracts the change curves of multiple geological parameters in the same area to calculate the time series correlation coefficient, determines the correlation anomaly interval based on the matching degree of the correlation coefficient and the directional offset difference, and obtains the distribution result of the anomaly correlation coefficient.

[0024] The causal chain generation submodule, based on the abnormal correlation coefficient distribution results, filters regions where the directional offset difference is opposite to the correlation direction, extracts spatial location and corresponding parameter offset features, adjusts the causal direction order, rearranges the dependencies between multiple parameters, and generates perturbed causal chain results.

[0025] As a further aspect of the present invention, the directional offset reference value is obtained by acquiring the geological parameter change sequence in multiple regions, calculating the directional vector between each pair of adjacent parameters, and setting the mean and variance of the included angle based on the change trend.

[0026] As a further aspect of the present invention, the resonance response module includes:

[0027] The data analysis submodule acquires time series data of multiple geological parameters within the disturbed area, detects sampling intervals and data gaps and fills in missing sections, then rearranges them according to sampling time, calculates the rate of change of multiple parameters between consecutive time points as the rate of disturbance change, and generates a set of geological parameter time change series.

[0028] The phase synchronization submodule, based on the geological parameter time change sequence set, identifies the triggering event point, extracts the instantaneous phase value near the event point, calculates the phase difference between parameters and counts the stability within the event interval, analyzes the average synchronization relationship strength between parameters, and generates the phase synchronization rate distribution result.

[0029] The resonance assessment submodule calls the phase synchronization rate distribution results, extracts the parameter group with the optimal phase synchronization rate, calculates the resonance intensity coefficient between multiple geological parameters, and uses the resonance intensity coefficient and disturbance amplitude as input to weightedly summarize and sort the influence intensity of the parameter combination to generate preliminary mineral exploration results.

[0030] As a further aspect of the present invention, the optimized mineral exploration module includes:

[0031] The resonance smoothing submodule extracts the resonance coefficient difference between adjacent intervals and constrains the amplitude of the difference based on the resonance coefficients in the same region of the preliminary prospecting results and the perturbation causal chain results, adjusts the resonance coefficient values ​​of multiple intervals and reconstructs the distribution sequence to generate a smooth resonance coefficient sequence.

[0032] The iterative correction submodule calls the smoothed resonance coefficient sequence, detects the distribution density of extreme values ​​in the sequence and identifies isolated outliers, replaces the outliers with the mean of adjacent intervals, recalculates the rate of change of the corrected sequence and iterates multiple times until the density is stable, and obtains the resonance distribution result after anomaly correction.

[0033] The mineralization scoring submodule extracts three indicators—disturbance response amplitude, phase synchronization rate, and resonance intensity coefficient—based on the resonance distribution results after anomaly correction. It calculates the weighted average of multiple indicators and synthesizes mineralization factors according to the weight ratio. It then analyzes and scores by region to generate optimized mineral exploration analysis results.

[0034] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0035] By acquiring and registering geological parameters in multiple dimensions, a unified expression of geological elements in time and space is achieved, effectively eliminating deviations under different observation conditions and ensuring comparability and continuity among data. Based on this, feature dimensionality reduction and mineralization parameter extraction are performed on the registered multi-source data, significantly reducing noise interference from data dimensionality and improving the accuracy of geological feature expression. Vectorized calculation of parameter gradients between adjacent sampling points enables quantifiable characteristics of the geological strain process, capturing subtle stress changes spatially and identifying potential tension concentration areas. By analyzing the directional offset and temporal correlation between multiple parameters, the true causal relationship of geological disturbances can be revealed, avoiding misjudgments caused by ambiguous correlations in traditional methods. Further analysis of phase synchronization and amplitude differences in temporal data of the disturbance area, combined with weighted calculation of resonance response, dynamically reflects the degree of response of the geological system under energy coupling. The overall logic realizes a hierarchical analysis from data fusion, spatial analysis, causal reasoning to dynamic response, giving mineral exploration results higher spatiotemporal resolution and interpretive credibility, improving the accuracy of mineralization information identification and the reliability of prediction. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a schematic diagram of the system of the present invention;

[0038] Figure 2 This is a schematic diagram of the system framework of the present invention;

[0039] Figure 3 This is a flowchart of the geological parameter module in this invention;

[0040] Figure 4 This is a flowchart of the spatial strain module in this invention;

[0041] Figure 5 This is a flowchart of the causal analysis module in this invention;

[0042] Figure 6 This is a flowchart of the resonance response module in this invention;

[0043] Figure 7 This is a flowchart of the optimized mineral exploration module in this invention. Detailed Implementation

[0044] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0045] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0046] This invention provides an artificial intelligence-based geological prospecting data processing and analysis system, such as... Figure 1-2 The diagram shown illustrates an artificial intelligence-based geological prospecting data processing and analysis system, which includes:

[0047] The geological parameter module acquires rock density gradient, geothermal gradient, fracture elongation, pore fluid pressure difference, and magnetic field changes and performs spatiotemporal registration. It then performs feature dimensionality reduction on the registered data and extracts the main mineralization parameters, generating a geological parameter set and transferring it to the spatial strain module.

[0048] The spatial strain module performs spatial distribution analysis based on the geological parameter set, calculates and quantifies the gradient difference of geological parameters between adjacent sampling points, analyzes the spatial strain intensity, and if it exceeds the spatial strain threshold, it extracts spatial tension points, generates tension distribution results, and transmits them to the causal analysis module.

[0049] The causal analysis module performs causal chain analysis based on the tension distribution results and geological parameter set, calculates the directional offset difference and temporal correlation coefficient between multiple geological parameters, filters out areas with reverse and abnormal correlation as disturbance areas and adjusts the causal direction, generates disturbance causal chain results and transmits them to the resonance response module.

[0050] The resonance response module acquires time-series data of multiple geological parameters within the disturbance area, identifies excitation event points, calculates phase synchronization rate and resonance coefficient by combining the results of the disturbance causal chain, evaluates the influence intensity of multiple geological parameters, and generates preliminary mineral exploration results.

[0051] The mineral exploration module is optimized by combining preliminary mineral exploration results with perturbation causal chain results, smoothing the distribution of resonance coefficients, eliminating isolated outliers through iterative updates, calculating mineralization factor scores, and generating optimized mineral exploration analysis results.

[0052] The geological parameter set includes rock density gradient, geothermal gradient, and fracture elongation rate. The tension distribution results include spatial strain intensity distribution and spatial tension point information. The disturbance causal chain results include directional offset difference, temporal correlation coefficient distribution, and disturbance area. The preliminary prospecting results include excitation event point set, phase synchronization rate distribution, and resonance coefficient. The optimized prospecting analysis results include smoothed resonance coefficient, ore-forming factor score, and elimination of isolated anomalies.

[0053] Specifically, such as Figure 2 , 3 As shown, the geological parameter module includes:

[0054] The geological data submodule acquires raw observation data on rock density gradient, geothermal gradient, fracture elongation, pore fluid pressure difference, and magnetic field changes; calculates the variation amplitude and interval average of multiple parameters within the same region; and performs stratigraphic calibration on the parameter sequence to generate the original geological parameter set.

[0055] Raw observational data on rock density gradient, geothermal gradient, fracture elongation, pore fluid pressure difference, and magnetic field changes are acquired through periodic data collection at multiple sensor stations deployed within the exploration area. In a specific implementation, taking a copper mine exploration area as an example, three monitoring points (numbered ZKD01, ZKD02, and ZKD03) are selected for data collection. Rock density gradient data is obtained through measurements at different depths using a microgravity meter; geothermal gradient data is measured using fiber optic temperature sensors embedded in the borehole; fracture elongation is calculated by monitoring surface displacement using differential GPS (DGPS); pore fluid pressure difference is measured using piezoresistive pressure gauges installed at different strata; and magnetic field changes are recorded by a proton magnetometer. Within one observation cycle (e.g., 24 hours), each parameter is measured multiple times to obtain a set of time-series data. For example, at monitoring point ZKD01, the rock density gradient sequence measured at a depth of 1200 meters is [0.21, 0.22, 0.21, 0.23] g / cm³ / km; the geothermal gradient sequence is [35.2, 35.5, 35.4, 35.8] °C / km. Subsequently, the variation amplitude and interval average of multiple parameters within the same region are calculated. The calculation process for the variation amplitude is as follows: extract the maximum and minimum values ​​of a parameter within a specified time interval and calculate the difference between them. Taking the rock density gradient sequence at point ZKD01 as an example, its maximum value is 0.23 g / cm³ / km, and its minimum value is 0.21 g / cm³ / km, so its variation amplitude is 0.23 - 0.21 = 0.02 g / cm³ / km. The calculation process for the interval average is as follows: sum all observed values ​​of a parameter within a specified time interval and then divide by the number of observations. The interval average of the rock density gradient sequence is (0.21+0.22+0.21+0.23) / 4 = 0.2175 g / cm³ / km. This calculation is performed on all five parameters at all monitoring points. Next, the parameter sequence is stratigraphically labeled. This process compares the calculated parameter changes with existing borehole core data to determine the specific geological strata corresponding to the parameter anomalies. For example, at the ZKD02 monitoring point, in the depth range of 1500-1550 meters, a geothermal gradient jump from 38°C / km to 45°C / km was observed, while borehole core data indicated that this depth is the contact zone between sandstone and granite intrusions. Therefore, this geothermal gradient jump is labeled at this contact zone stratum.

[0056] Table 1. Examples of geological parameter stratigraphic identification.

[0057] Depth (meters) stratigraphy Geothermal gradient (°C / km) Magnetic field change (nT) 1450-1500 sandstone 38.1 55 1500-1550 Contact strip 45.3 150 1550-1600 granite 42.5 120

[0058] As shown in Table 1, stratigraphic calibration was completed by associating each parameter value with known stratigraphic lithology. The original observation data from all monitoring points, the calculated variation range values, the interval average values, and the stratigraphic calibration information were integrated to form a structured dataset, namely the original geological parameter set.

[0059] The spatial registration submodule, based on the original geological parameter set, calls time series and spatial coordinate information to register multiple parameters, calculates the spatial offset and time delay difference of parameters between multiple observation points, and performs coordinate alignment and interpolation correction on multi-source data to generate a spatiotemporal registration parameter set;

[0060] Based on the generated original geological parameter set, the time series and spatial coordinate information recorded within it are used to register multiple parameters. The original geological parameter set contains the three-dimensional spatial coordinates (X, Y, Z) of each monitoring point (e.g., ZKD01, ZKD02, ZKD03) and the timestamp of each observation. For example, the coordinates of ZKD01 are (1523.4, 2845.1, -1200.0) meters, and the coordinates of ZKD02 are (1583.4, 2875.1, -1250.0) meters, with timestamps accurate to the second. Next, the spatial offset and time delay difference of the parameters between multiple observation points are calculated. The spatial offset is the straight-line distance between two monitoring points in three-dimensional space. Taking ZKD01 and ZKD02 as examples, their spatial offset is calculated as: the square root of (1583.4-1523.4)²+(2875.1-2845.1)²+(-1250.0-(-1200.0)²), which is the square root of 60²+30²+(-50)², resulting in the square root of (3600+900+2500), or 77.46 meters. The time delay difference is calculated by cross-correlation analysis of the time series with the same parameters at the two monitoring points. For example, analyzing ZKD01 and ZKD02... For the geothermal gradient time series of ZKD02, the ZKD02 series is shifted relative to the ZKD01 series for each time step (e.g., 1 hour), and the sum of the products of corresponding points in the two series is calculated after each shift. When the sum of the products reaches its maximum value, the corresponding shift amount is the time delay difference. If the sum of the products is maximum after shifting for 3 time steps, and the time step is 1 hour, then the time delay difference is 3 hours. Then, coordinate alignment and interpolation correction are performed on the multi-source data. First, a unified three-dimensional grid coordinate system is established, with the grid cell size based on the exploration precision. The required distance is set, for example, 10 meters × 10 meters × 10 meters. Since the original monitoring points (ZKD01, ZKD02, etc.) may not fall exactly on the grid nodes, interpolation correction is required. The interpolation process uses an inverse distance weighting method: for any grid node, its parameter value is calculated based on all original monitoring points within a certain distance range around it. The weighting coefficient is set according to the reciprocal of the distance between the monitoring point and the grid node. For example, to calculate the pore fluid pressure difference at a certain grid node G, its distance from ZKD01 is 25 meters (pressure difference is 5.2 MPa). a) The distance to ZKD02 is 40 meters (pressure difference is 5.8 MPa). The weight of ZKD01 is 1 / 25 = 0.04, and the weight of ZKD02 is 1 / 40 = 0.025. The normalized weights are W1 = 0.04 / (0.04 + 0.025) = 0.615 and W2 = 0.025 / (0.04 + 0.025) = 0.385. Therefore, the estimated pressure difference at node G is 5.2 × 0.615 + 5.8 × 0.385 = 3.198 + 2.233 = 5.431 MPa.By performing this interpolation calculation on all grid nodes, discrete monitoring data is transformed into a continuous, regular data field that fills the entire three-dimensional space of the exploration area, thereby generating a spatiotemporal registration parameter set.

[0061] The parameter extraction submodule performs feature dimensionality reduction on rock density gradient, geothermal gradient, fracture elongation rate, pore fluid pressure difference and magnetic field changes based on the spatiotemporal registration parameter set, calculates the contribution rate of multiple parameters in the principal components of mineralization, extracts key mineralization feature sequences, and generates a geological parameter set.

[0062] Based on the generated spatiotemporal registration parameter set, feature reduction is performed on five parameters: rock density gradient, geothermal gradient, fracture elongation, pore fluid pressure difference, and magnetic field variation. This process first constructs an N×5 matrix by calculating the values ​​of these five parameters at all grid nodes, where N is the total number of grid nodes. Then, a 5×5 covariance matrix is ​​calculated based on this matrix. The diagonal elements of the covariance matrix represent the variance of each parameter itself, while the off-diagonal elements represent the pairwise covariance between parameters. Next, the eigenvalues ​​and eigenvectors of this covariance matrix are calculated. For example, the five calculated eigenvalues ​​are λ1=8.5, λ2=3.2, λ3=0.8, λ4=0.2, and λ5=0.1. Then, the contribution rate of each parameter to the principal components of ore formation is calculated. The contribution rate of each principal component is determined by the proportion of its corresponding eigenvalue to the sum of the total eigenvalues. The sum of the total eigenvalues ​​is 8.5+3.2+0.8+0.2+0.1=12.8. The contribution rate of the first principal component (i.e., the ore-forming principal component) is 8.5 / 12.8 = 66.4%. The contribution rate of the second principal component is 3.2 / 12.8 = 25.0%. A cumulative contribution rate threshold is set to select the principal components with the largest contributions. This threshold is determined by backtracking analysis of historical data from known mining areas. Experimental data shows that when the cumulative contribution rate reaches 90%, the included principal components can explain most of the data variation and have the highest consistency with the ore-forming enrichment areas. In this example, the cumulative contribution rate of the first two principal components is 66.4% + 25.0% = 91.4%, which is greater than the 90% threshold, so the first two principal components are selected. Among them, the first principal component with the largest eigenvalue is defined as the ore-forming principal component. Next, key ore-forming feature sequences are extracted. This process first calculates the score of each grid node on the ore-forming principal component. This score is the sum of the products of the standardized values ​​of the five parameters at that node and the corresponding eigenvector of the principal component. Then, a feature extraction threshold is set to select nodes with high scores. The threshold is set to the mean score of all grid nodes plus twice the standard deviation. This setting is based on statistical analysis of historical data, which found that the probability of ore bodies existing in areas exceeding this threshold is higher than 95%. For example, if the mean score of all nodes is 0.5 and the standard deviation is 1.2, then the threshold is 0.5 + 2 × 1.2 = 2.9. All grid nodes with scores higher than 2.9 are identified, and the spatial coordinate sequences of these nodes and their corresponding parameter values ​​together constitute the key mineralization feature sequence. Finally, this sequence is integrated with the selected principal component information to generate a set of geological parameters.

[0063] Specifically, such as Figure 2 , 4 As shown, the space strain module includes:

[0064] The gradient difference calculation submodule performs spatial distribution analysis based on the geological parameter set, extracts the coordinate positions and parameter value sequences of adjacent sampling points, calculates the difference values ​​of geological parameters between multiple points, and standardizes the difference results according to the spatial distance weight to generate a spatial parameter gradient difference set.

[0065] Spatial distribution analysis based on a geological parameter set is performed by first extracting the coordinates and parameter value sequences of adjacent sampling points within the key mineralization feature sequences from the dataset. Taking two adjacent grid nodes as an example, node A has coordinates of (1600, 2900, -1550) meters, and node B has coordinates of (1610, 2900, -1550) meters. The five geological parameter values ​​corresponding to these two nodes are retrieved from the geological parameter set, as shown in Table 2.

[0066] Table 2 Geological parameter values ​​of adjacent nodes

[0067] Parameter name Node A parameter value Node B parameter value Rock density gradient 0.25 0.28 geothermal gradient 46.1 46.5 Elongation at break 0.0015 0.0017 Pore ​​fluid pressure difference 6.1 6.3 Magnetic field changes 165 172

[0068] As shown in Table 2, detailed parameter values ​​for nodes A and B were obtained. Next, the differences in geological parameters between multiple points were calculated. The difference was calculated by subtracting the parameter value of node A from the parameter value of node B. Taking the rock density gradient as an example, its difference is 0.28 - 0.25 = 0.03. The difference in geothermal gradient is 46.5 - 46.1 = 0.4. The difference in fracture elongation is 0.0017 - 0.0015 = 0.0002. The difference in pore fluid pressure difference is 6.3 - 6.1 = 0.2. The difference in magnetic field variation is 172 - 165 = 7. Subsequently, the difference results were standardized according to spatial distance weights. First, the spatial distance between two points was calculated as the square root of the sum of the squares of the differences in the three-dimensional coordinates of the two points. The spatial distance between node A and node B is the square root of ((1610-1600)² + (2900-2900)² + (-1550-(-1550))², which is 10 meters. The normalization process converts the difference values ​​into the rate of change per unit distance, i.e., the gradient. The gradient of the rock density gradient is 0.003 g / cm³ / km / m. The gradient of the geothermal gradient is 0.04°C / km / m. This normalization process does not introduce additional weighting coefficients; instead, it normalizes the difference values ​​to gradient values ​​by dividing by the actual distance. This calculation is repeated for all adjacent node pairs, and the resulting gradient values ​​for each parameter are summarized to generate a set of spatial parameter gradient differences.

[0069] The intensity analysis submodule calls the spatial parameter gradient difference set, calculates the direction vector amplitude of the geological parameter gradient in each region, analyzes the strain intensity value in the region based on the amplitude variation interval, and combines the mean variation of intensity in adjacent regions to generate spatial strain intensity distribution results.

[0070] The generated spatial parameter gradient difference set is used to calculate the magnitude of the direction vector of the geological parameter gradient in each region. A region consists of a set of adjacent nodes in a 3D mesh, and the gradient vector of the region is determined by the average gradient calculated for all nodes within that region. Taking a region consisting of a center point and its six adjacent points in all directions as an example, the average gradient vector of its five geological parameters is [0.003, 0.04, 0.00002, 0.02, 0.7], with units of g / cm³ / km / m, °C / km / m, m⁻¹, MPa / m, and / m, respectively. Before calculating the magnitude, to eliminate the influence of different parameter dimensions, each gradient component needs to be normalized. The normalization coefficient is set based on statistical analysis of a large amount of geological survey data from similar mining areas, selecting the 95th percentile of the gradient value of each parameter as the benchmark. Analysis yielded the following normalized baseline values ​​for the five parameters: 0.01 g / cm³ / km / m, 0.1°C / km / m, 0.0001 m⁻¹, 0.05 MPa / m, and 2.0 m. The normalized gradient vector is then [0.003 / 0.01, 0.04 / 0.1, 0.00002 / 0.0001, 0.02 / 0.05, 0.7 / 2.0] = [0.3, 0.4, 0.2, 0.4, 0.35]. The magnitude of the direction vector in this region is the Euclidean norm of this normalized vector, calculated as the square root of (0.3² + 0.4² + 0.2² + 0.4² + 0.35²), which is the square root of (0.09 + 0.16 + 0.04 + 0.16 + 0.1225), resulting in 0.757. The strain intensity values ​​within the region are analyzed based on the amplitude variation interval. The criteria for dividing the strain intensity interval are determined by statistical cluster analysis of the amplitude values ​​of over 1000 sample regions in known mineralized and unmineralized areas. Strain intensity is divided into three levels: low strain intensity region (amplitude values ​​between 0.0 and 0.5), medium strain intensity region (amplitude values ​​between 0.5 and 1.5), and high strain intensity region (amplitude values ​​greater than 1.5). In this example, the calculated amplitude value of 0.757 belongs to the medium strain intensity region. Next, the analysis is combined with the mean variation of the intensity in adjacent regions. The calculated direction vector amplitude of another region adjacent to the current region is set to 1.62, belonging to the high strain intensity region. The mean variation of the intensity between the two regions is |1.62 - 0.757| = 0.863. This variation value is sorted with the intensity variation values ​​of all adjacent region pairs to identify the boundary of the region with the most drastic change. The strain intensity values ​​and spatial location information of each region are integrated to generate the spatial strain intensity distribution results.

[0071] The tension point extraction submodule identifies regions exceeding the spatial strain threshold based on the spatial strain intensity distribution results, extracts the center coordinates and corresponding intensity indices as a set of tension points, calculates the spatial distance and directional distribution of adjacent tension points, and generates spatial tension distribution results.

[0072] Based on the generated spatial strain intensity distribution results, areas exceeding the spatial strain threshold are identified. The spatial strain threshold is set with reference to the statistical analysis of strain intensity values ​​in areas where known high-grade ore bodies are located. Twenty typical deposits were selected, and detailed calculations of the spatial strain intensity distribution within and around them were performed. Experimental data showed that 95% of the high-grade ore bodies are located in areas with strain intensity values ​​greater than 1.55. Therefore, the spatial strain threshold is set to 1.55. The intensity values ​​of all areas in the spatial strain intensity distribution results are compared with 1.55. In the aforementioned example, the strain intensity value of a certain area is 1.62. Since 1.62 is greater than 1.55, this area is identified as a high-strain zone. The center coordinates of this area and the corresponding intensity index are extracted as an element in the tension point set. The center coordinates of the area are obtained by calculating the arithmetic mean of the three-dimensional coordinates of all grid nodes within the area. The high-strain zone contains three nodes with coordinates (1810, 3150, -1600), (1820, 3150, -1600), and (1810, 3160, -1600). The center coordinates are (1810+1820+1810) / 3, (3150+3150+3160) / 3, and (-1600-1600-1600) / 3, which equals (1813.3, 3153.3, -1600.0). Therefore, a tension point T1 is generated with coordinates (1813.3, 3153.3, -1600.0) and a strength of 1.62. By screening the entire exploration area, a set of tension points is obtained. For example, there is another tension point T2 with coordinates (1853.3, 3183.3, -1620.0) and intensity of 1.71. Then, the spatial distance and directional distribution of adjacent tension points are calculated. The spatial distance between T1 and T2 is the square root of (1853.3-1813.3)²+(3183.3-3153.3)²+(-1620.0-(-1600.0))², which is the square root of 40²+30²+(-20)², resulting in the square root of (1600+900+400), or 53.85 meters. The direction vector from T1 to T2 is 1853.3-1813.3, 3183.3-3153.3, -1620.0-(-1600.0), which is (40, 30, -20). Normalizing this vector yields the direction distribution information. Summarizing all tension points and their spatial distances and directions generates the spatial tension distribution result.

[0073] Specifically, such as Figure 2 , 5 As shown, the causal analysis module includes:

[0074] The offset difference calculation submodule, based on the tension distribution results and the spatial coordinate sequence of the geological parameter set, detects the direction vectors of adjacent geological parameters and calculates the change in the angle between multiple parameters. Combined with the direction offset benchmark value, it judges the degree of direction difference between multiple regions and generates a direction offset difference sequence.

[0075] Based on the tension distribution results and the spatial coordinate sequence of the geological parameter set, the regions containing the two tension points extracted in the previous steps, T1 (1813.3, 3153.3, -1600.0) and T2 (1853.3, 3183.3, -1620.0), are analyzed. First, the direction vectors of adjacent geological parameters within these two regions are detected and extracted from the geological parameter set. Here, the direction vector refers to the gradient vector of each parameter in three-dimensional space. Taking the geothermal gradient and magnetic field change parameters as examples, the average gradient vectors of regions T1 and T2 are retrieved from the spatial parameter gradient difference set. In region T1, the geothermal gradient vector is [0.04, 0.03, -0.01], and the magnetic field change gradient vector is [0.7, 0.5, -0.2]. In region T2, the geothermal gradient vector is [0.05, 0.02, -0.015], and the magnetic field change gradient vector is [0.6, 0.8, -0.1]. Subsequently, the angle variation between multiple parameters is calculated. In region T1, the dot product of the two gradient vectors is 0.04×0.7+0.03×0.5+(-0.01)×(-0.2)=0.028+0.015+0.002=0.045. The magnitude of the geothermal gradient vector is (0.04²+0.03²+(-0.01)²)^0.5=0.051, and the magnitude of the magnetic field change gradient vector is (0.7²+0.5²+(-0.2)²)^0.5=0.883. The angle in region T1 is arccos(0.045 / (0.051×0.883))=arccos(0.999)≈2.5 degrees. In region T2, the dot product of the two gradient vectors is 0.05×0.6+0.02×0.8+(-0.015)×(-0.1)=0.03+0.016+0.0015=0.0475. The magnitude of the geothermal gradient vector is 0.056, and the magnitude of the magnetic field gradient vector is 1.005. The angle between the two gradient vectors in region T2 is arccos(0.0475 / (0.056×1.005))=arccos(0.844)≈32.4 degrees. The change in angle is |32.4-2.5|=29.9 degrees. Next, the degree of directional difference between multiple regions is determined by combining the directional offset benchmark value. The directional offset benchmark value is determined by analyzing 50 known non-mineralized stable geological structures. The statistical results show that the average change in the angle between the geothermal gradient and the magnetic field gradient vectors is 5.1 degrees, and the standard deviation is 1.7 degrees. The baseline value was set as the mean plus three standard deviations, i.e., 5.1 + 3 × 1.7 = 10.2 degrees. Since the calculated angle change of 29.9 degrees is much greater than the baseline value of 10.2 degrees, the directional difference between regions T1 and T2 was determined to be "high". This calculation was then extended to all adjacent tension point pairs to generate a sequence of directional offset differences.

[0076] The time series assessment submodule calls the time series data of the directional offset difference sequence and the geological parameter set, extracts the change curves of multiple geological parameters in the same area to calculate the time series correlation coefficient, determines the correlation anomaly interval based on the matching degree of the correlation coefficient and the directional offset difference, and obtains the distribution results of the anomaly correlation coefficient.

[0077] The generated directional offset difference sequence and time series data from the geological parameter set were used to analyze region T1. The directional offset difference in this region is 29.9 degrees, which is considered a high degree of difference. Next, the variation curves of geothermal gradient and magnetic field changes in this region were extracted over 10 consecutive time steps (each step is 6 hours), and the specific data are shown in Table 3.

[0078] Table 3. Time series data of geological parameters in the T1 area

[0079] Time step geothermal gradient Magnetic field changes 1 46.1 165 2 46.3 168 3 46.2 166 4 46.5 170 5 46.4 169 6 46.7 173 7 46.8 175 8 46.6 172 9 47.0 178 10 46.9 176

[0080] As shown in Table 3, time series of the two parameters were obtained. Then, the temporal correlation coefficient between these two series was calculated. First, the average values ​​of the two series were calculated: the average geothermal gradient was 46.61°C / km, and the average magnetic field change was 171.2 nT. Then, the deviation of each data point from the average value was calculated, and the corresponding deviations were multiplied and summed to obtain a covariance correlation term of 4.69. Next, the standard deviations of the two series were calculated: the standard deviation of the geothermal gradient was 0.273, and the standard deviation of the magnetic field change was 4.44. The temporal correlation coefficient (Pearson correlation coefficient) was 4.69 / (0.273×4.44)=4.69 / 1.21=0.98. The correlation anomaly intervals were determined based on the matching degree between the correlation coefficient and the directional offset difference. The matching degree rule is based on the following premise: under normal stress transmission, regions with large differences in spatial gradient direction (high directional offset difference) should have poor temporal evolution synchronization of their internal parameters (low correlation coefficient). The specific criteria are as follows: regions with a directional offset difference greater than 10.2 degrees and a correlation coefficient greater than 0.5 are considered to have abnormal correlation. In this example, the directional offset difference in region T1 is 29.9 degrees (high), while the temporal correlation coefficient is 0.98 (high). This result does not meet the preset matching relationship, therefore region T1 is identified as an abnormal correlation interval. All identified abnormal intervals and their correlation coefficients are summarized to obtain the abnormal correlation coefficient distribution results.

[0081] The causal chain generation submodule filters regions whose directional offset difference is opposite to the direction of correlation based on the abnormal correlation coefficient distribution results, extracts spatial location and corresponding parameter offset features, adjusts the causal direction order, rearranges the dependencies between multiple parameters, and generates perturbation causal chain results.

[0082] Based on the generated distribution of anomalous correlation coefficients, region T1 was selected for in-depth analysis. This region satisfies the condition that "the direction of the offset difference is opposite to the direction of the correlation," i.e., the direction offset difference is 29.9 degrees (higher than the threshold of 10.2 degrees), and the temporal correlation coefficient is 0.98 (also higher than the threshold of 0.5). Next, the spatial location and corresponding parameter offset features of this anomalous region were extracted. The spatial location is the center coordinates of T1 (1813.3, 3153.3, -1600.0). The parameter offset features are the geothermal gradient vector [0.04, 0.03, -0.01] °C / km / m and the magnetic field change gradient vector [0.7, 0.5, -0.2] nT / m. Subsequently, the causal direction order was adjusted, and the dependencies between multiple parameters were rearranged. In a conventional geological environment, the causal relationship between geological parameters is usually manifested as stress field changes (characterized by parameters such as fracture elongation) causing changes in rock pore structure, which in turn affect fluid pressure and geothermal field, ultimately leading to corresponding changes in the magnetic field. The dependency can be simplified as: stress → geothermal temperature → magnetic field. However, in region T1, geothermal temperature and magnetic field show significant differences in spatial gradient direction, but exhibit strong temporal synchronicity (correlation coefficient 0.98). This indicates that the two are not simply sequentially causally related, but are synchronously driven by a common and powerful deep disturbance source. Therefore, the dependency in this region is rearranged, introducing an unobserved "deep hydrothermal activity" as the top-level driving factor. The adjusted dependency is: "deep hydrothermal activity" simultaneously causes "geothermal gradient anomalies" and "magnetic field anomalies." This new dependency, along with its corresponding spatial location and parameter characteristics, forms the disturbance causal chain for region T1. This analysis is performed on all selected anomaly regions to generate the disturbance causal chain results.

[0083] Specifically, such as Figure 2 , 6 As shown, the resonance response module includes:

[0084] The data analysis submodule acquires time series data of multiple geological parameters within the disturbed area, detects sampling intervals and data gaps and fills in missing sections, then rearranges them according to sampling time, calculates the rate of change of multiple parameters between consecutive time points as the rate of disturbance change, and generates a set of geological parameter time change series.

[0085] Time series data for two geological parameters, geothermal gradient and magnetic field variation, were acquired within region T1 (center coordinates: 1813.3, 3153.3, -1600.0). This data consists of 10 data points spanning 60 consecutive hours with a sampling interval of 6 hours. During data analysis, a complete geothermal gradient sequence was found, but a gap was observed in the magnetic field variation sequence at time point 8 (48 hours). This missing segment was filled using linear interpolation. The magnetic field variation value at time point 7 was 175 nanotesla, and at time point 9 it was 178 nanotesla. Therefore, the filled value for time point 8 was (175 + 178) divided by 2, resulting in 176.5 nanotesla. After filling, the two complete time series containing geothermal gradient and magnetic field variation data were rearranged according to sampling time. Subsequently, the rate of change of multiple parameters between consecutive time points was calculated; this is the perturbation rate of change. The rate of change was calculated by subtracting the value of the previous time point from the value of the next time point, and then dividing by the sampling interval of 6 hours. Taking the geothermal gradient sequence as an example, the rate of change at the second time point (12 hours) is (46.3-46.1) / 6 = 0.033 degrees Celsius per kilometer per hour. The rate of change at the third time point is (46.2-46.3) / 6 = -0.017 degrees Celsius per kilometer per hour. Performing this calculation on the entire time series of geothermal gradient and magnetic field changes yields two rate of change sequences of length 9. For example, the rate of change of the magnetic field at the second time point is (168-165) / 6 = 0.5 nanoteslas per hour. Integrating these two rate of change sequences generates a set of geological parameter time variation sequences.

[0086] The phase synchronization submodule, based on the time variation sequence set of geological parameters, identifies the triggering event point, extracts the instantaneous phase value near the event point, calculates the phase difference between parameters and statistically analyzes the stability within the event interval, analyzes the average synchronization relationship strength between parameters, and generates the phase synchronization rate distribution result.

[0087] Based on the generated set of temporal variation sequences of geological parameters, triggering event points are first identified within the sequences. The identification criterion for triggering event points is that the absolute value of the rate of change of any parameter first exceeds the mean of its sequence plus twice the standard deviation. Taking the geothermal gradient rate of change sequence as an example, its mean is 0.016 degrees Celsius per kilometer per hour, and its standard deviation is 0.035. The threshold is set to 0.016 + (2 * 0.035) = 0.086. At the 7th time point, the geothermal gradient rate of change is (46.8 - 46.7) / 6 = 0.017, which does not exceed the threshold; however, at the 9th time point, the rate of change is (47.0 - 46.6) / 6 = 0.067, which also does not exceed the threshold. The magnetic field rate of change sequence is then re-examined; its mean is 1.28 nanoteslas per hour, and its standard deviation is 0.89. The threshold is 1.28 + (2 * 0.89) = 3.06. At time point 9, the magnetic field change rate was (178-172) / 6 = 1.0, which did not exceed the threshold. The criteria for identifying the excitation event point were adjusted, defining it as the maximum value point in the rate of change sequence, i.e., the peak value of the rate of change of the geothermal gradient near time point 9. Next, the instantaneous phase values ​​near the event point were extracted. This process involved performing a Hilbert transform on the time series to obtain the phase angle at each time point. At time point 9, the instantaneous phase value of the geothermal gradient sequence was 1.52 radians, and the instantaneous phase value of the magnetic field change sequence was 1.45 radians. The phase difference between the two parameters was calculated as |1.52-1.45|, or 0.07 radians. Further analysis of the stability of the phase difference within the event interval (time points 7 to 10) was performed. By calculating the standard deviation of the phase difference sequence [0.10, 0.08, 0.07, 0.09] within this interval, a value of 0.0129 was obtained. This value is less than the stability benchmark value of 0.05 obtained by performing the same analysis on 30 sets of background noise data, and is therefore considered stable. Finally, the average synchronization strength between the parameters was analyzed, calculated by subtracting the circumferential variance of the phase difference within the event interval from 1, and the result was 0.96. The synchronization strength value of 0.96 for the geothermal gradient and magnetic field change parameter pair in region T1 was recorded and used together with the calculation results of other parameter pairs to generate the phase synchronization rate distribution result.

[0088] The resonance assessment submodule calls the phase synchronization rate distribution results, extracts the parameter group with the optimal phase synchronization rate, calculates the resonance intensity coefficient between multiple geological parameters, and uses the resonance intensity coefficient and disturbance amplitude as input to weightedly summarize and sort the influence intensity of the parameter combination to generate preliminary mineral exploration results.

[0089] The generated phase synchronization rate distribution results are retrieved, and the parameter set with the optimal phase synchronization rate within the T1 region is extracted. Comparison shows that the geothermal gradient and magnetic field variation parameter set has the highest phase synchronization rate, with a value of 0.96. Next, the resonance intensity coefficient between these two geological parameters is calculated. This coefficient is obtained by multiplying the phase synchronization rate by a frequency consistency factor. The frequency consistency factor is obtained by performing a Fourier transform on the time series of the two parameters and calculating the product of their dominant frequency energy proportions. The calculated frequency consistency factor is 0.92 for the geothermal gradient sequence and 0.95 for the magnetic field variation sequence, resulting in a combined value of 0.87. Therefore, the resonance intensity coefficient is 0.96 multiplied by 0.87, resulting in 0.835. Then, using the resonance intensity coefficient and disturbance amplitude as inputs, the influence intensity of the parameter combination is weighted and summarized. The disturbance amplitude value is defined as the ratio of the maximum value of the parameter within the observation period to the average value of the parameter across the entire exploration area. The maximum value of the geothermal gradient is 47.0°C / km, the regional average is 39.5°C / km, and its disturbance amplitude value is 1.19. The maximum magnetic field variation was 178 nT, with a regional average of 120 nT and a disturbance amplitude of 1.48. The average disturbance amplitude of the parameter combination was (1.19 + 1.48) / 2 = 1.335. The weighting coefficients were set based on a retrospective verification experiment of 40 known mineral deposit samples, which showed that the resonance intensity coefficient had a higher correlation with mineralization potential. Through sensitivity analysis, the weight of the resonance intensity coefficient was determined to be 0.65, and the weight of the disturbance amplitude value was determined to be 0.35. Therefore, the total influence intensity of region T1 was (0.835 * 0.65) + (1.335 * 0.35) = 1.01. This calculation was applied to all identified disturbance regions, and the results were sorted.

[0090] Table 4 Ranking of the Intensity of the Disturbance Area

[0091] Area code Total score of influence intensity Mineral exploration potential ranking T1 1.01 1 T5 0.89 2 T3 0.75 3

[0092] As shown in Table 4, region T1 has the highest total score for influence intensity, ranking first in mineral exploration potential. This ranking list represents the preliminary mineral exploration results.

[0093] Specifically, such as Figure 2 , 7 As shown, the optimized mineral exploration module includes:

[0094] The resonance smoothing submodule extracts the resonance coefficient difference between adjacent intervals and constrains the amplitude of the difference based on the resonance coefficients in the same region in the preliminary prospecting results and the perturbation causal chain results. It then adjusts the resonance coefficient values ​​of multiple intervals and reconstructs the distribution sequence to generate a smooth resonance coefficient sequence.

[0095] Based on the resonance coefficients of the same region in the preliminary prospecting results and the perturbation causal chain results, the regions with the highest prospecting potential are processed. Region T1 (ranked first) and its adjacent region T5 (ranked second) are extracted from the preliminary prospecting results, and their resonance intensity coefficients are obtained from the perturbation causal chain results, which are 0.835 and 0.710, respectively. First, the difference in resonance coefficients between adjacent regions is extracted, i.e., the absolute value of the difference in resonance intensity coefficients between regions T1 and T5 is calculated: |0.835-0.710|=0.125. Then, an amplitude constraint is applied to this difference. The baseline value for the amplitude constraint is determined through resonance coefficient measurement experiments on 100 regions with continuous geological structures and no large fault structures. Experimental data shows that the resonance coefficient difference between 95% of adjacent regions is less than 0.10. Therefore, the threshold for the amplitude constraint is set to 0.10. Since the calculated difference of 0.125 is greater than this threshold, adjustment is required. The resonance coefficient values ​​across multiple intervals are adjusted as follows: The difference exceeding the threshold, i.e., 0.125, is subtracted by 0.10, resulting in 0.025. This 0.10 is then evenly distributed between two adjacent intervals. The distribution amount is 0.025 / 2 = 0.0125. For the T1 interval, which has a higher value, its resonance coefficient is lowered, resulting in a value of 0.835 - 0.0125 = 0.8225. For the T5 interval, which has a lower value, its resonance coefficient is raised, resulting in a value of 0.710 + 0.0125 = 0.7225. This operation smooths out the previously abrupt coefficient differences. This smoothing operation is applied to all adjacent interval pairs in the mineral potential ranking. For example, if the difference between the T5 and T3 intervals (resonance coefficient of 0.680) is |0.710 - 0.680| = 0.030, which is less than the threshold of 0.10, then the resonance coefficients between the T5 and T3 intervals are not adjusted in this step. The smoothed resonance coefficient values ​​of all regions are reconstructed into a distribution sequence according to their spatial location to generate a smooth resonance coefficient sequence.

[0096] The iterative correction submodule calls the smoothed resonance coefficient sequence, detects the distribution density of extreme values ​​in the sequence and identifies isolated outliers, replaces the outliers with the mean of adjacent intervals, recalculates the rate of change of the corrected sequence and iterates multiple times until the density is stable, and obtains the resonance distribution result after anomaly correction.

[0097] The generated smooth resonance coefficient sequence is called and iteratively corrected. The smooth resonance coefficient sequence of five consecutive intervals obtained along a certain exploration profile is: [0.823, 0.795, 0.350, 0.788, 0.760]. First, the distribution density of extreme values ​​in the sequence is detected and isolated anomalies are identified. The identification method is as follows: for each point in the sequence, the absolute value of the difference between it and its two adjacent points is calculated. If both of these differences are greater than a preset anomaly judgment threshold, the point is identified as an isolated anomaly. This threshold is set based on statistical analysis of 200 sets of simulation data, and three times the standard deviation of the difference between adjacent points in a sequence without abrupt changes is used as the threshold, which is set to 0.20. For the third point 0.350 in the sequence, the absolute value of the difference between it and the previous point 0.795 is |0.350-0.795|=0.445; the absolute value of the difference between it and the next point 0.788 is |0.350-0.788|=0.438. Since both 0.445 and 0.438 are greater than the threshold of 0.20, point 0.350 is identified as an isolated outlier. Subsequently, this outlier is corrected by replacing the value with the mean of adjacent intervals. The value of point 0.350 is replaced with the arithmetic mean of its two preceding and following points (0.795 and 0.788), calculated as (0.795 + 0.788) / 2 = 0.7915. The corrected sequence is: [0.823, 0.795, 0.7915, 0.788, 0.760]. Next, the rate of change of the corrected sequence is recalculated and iterated until the density stabilizes. The rate of change refers to the difference between adjacent points in the sequence. The original sequence's rate of change sequence is [-0.028, -0.445, 0.438, -0.028]. The corrected sequence's rate of change is [-0.028, -0.0035, -0.0035, -0.028]. The criterion for density stability is that the standard deviation of the rate of change is less than 0.05. After the first iteration, the standard deviation of the rate of change sequence is 0.012, which is less than 0.05, therefore the iteration stops. The corrected sequence is taken as the result of the resonance distribution after anomaly correction.

[0098] The mineralization scoring submodule extracts three indicators—disturbance response amplitude, phase synchronization rate, and resonance intensity coefficient—based on the resonance distribution results after anomaly correction. It calculates the weighted average of multiple indicators and synthesizes mineralization factors according to the weight ratio. It then analyzes and scores by region to generate optimized mineral exploration analysis results.

[0099] Based on the generated resonance distribution results after anomaly correction, the mineralization score of region T1 was determined. First, three key indicators for region T1 were extracted. The first indicator was the disturbance response amplitude, with a value of 1.335, obtained by taking the arithmetic mean of the disturbance amplitudes of the geothermal gradient and magnetic field changes (1.19 and 1.48, respectively). The second indicator was the phase synchronization rate, with a value of 0.96, derived by calculating the stability of the phase difference between the geothermal gradient and magnetic field changes within the event interval. The third indicator was the resonance intensity coefficient after anomaly correction, with a value of 0.823, obtained after resonance smoothing and iterative correction. Next, the weighted average of these three indicators was calculated to synthesize the mineralization factor for the region. The weighting coefficients were set based on multiple regression analysis of various indicators and actual reserves of 50 proven mineral deposits. The analysis results showed that the resonance intensity coefficient had the highest explanatory power, followed by the phase synchronization rate, while the disturbance response amplitude was relatively low. Accordingly, weights were assigned to the three indicators: the resonance intensity coefficient was weighted at 0.45, the phase synchronization rate at 0.35, and the disturbance response amplitude at 0.20. The sum of these weights was 1.0. The calculation process for the mineralization factor in region T1 was: (1.335*0.20)+(0.96*0.35)+(0.823*0.45)=0.973. Then, based on the regional analysis score, this mineralization factor was used as the final score for that region. This calculation was repeated for all potential regions, and they were ranked according to their scores.

[0100] Table 5. Ranking of Mineral Exploration Potential After Optimization

[0101] Area code Mineralization factor score Mineral exploration potential ranking T1 0.973 1 T5 0.861 2 T3 0.782 3

[0102] As shown in Table 5, the mineralization factor score for region T1 is 0.973, ranking first among all regions. This score ranking list represents the optimized mineral exploration analysis results.

[0103] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An artificial intelligence-based geological prospecting data processing and analysis system, characterized in that, The system comprises: a geological parameter module, which acquires rock density gradient, geothermal gradient, fracture extension rate, pore fluid pressure difference and magnetic field change and performs space-time registration, performs feature dimension reduction on the registered data and extracts main ore-forming parameters, generates a geological parameter set and delivers it to a spatial strain module; the spatial strain module, which performs spatial distribution analysis based on the geological parameter set, calculates the gradient difference of geological parameters between adjacent sampling points and vectorizes, analyzes the spatial strain intensity, extracts the spatial tension points if the spatial strain threshold is exceeded, generates a tension distribution result and delivers it to a causal analysis module; the causal analysis module, which performs causal chain analysis based on the tension distribution result and the geological parameter set, calculates the directional offset difference and time series correlation between multiple geological parameters, screens the reverse and abnormally correlated areas as disturbance areas and adjusts the causal direction, generates a disturbance causal chain result and delivers it to a resonance response module; the resonance response module, which acquires the time series data of multiple geological parameters in the disturbance area, identifies the excitation event points and calculates the phase synchronization rate and resonance coefficient in combination with the disturbance causal chain result, and evaluates the influence intensity of multiple geological parameters, generates a preliminary prospecting result.

2. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 1, characterized in that: The geological parameter set comprises rock density gradient, geothermal gradient, fracture extension rate, the tension distribution result comprises spatial strain intensity distribution and spatial tension point information, the disturbance causal chain result comprises directional offset difference, time series correlation distribution and disturbance area, and the preliminary prospecting result comprises excitation event point set, phase synchronization rate distribution and resonance coefficient.

3. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 1, characterized in that, The geological parameter module comprises: a geological data submodule, which acquires the original observation data of rock density gradient, geothermal gradient, fracture extension rate, pore fluid pressure difference and magnetic field change, calculates the change amplitude value and interval average value of multiple parameters in the same area, and performs horizon calibration on the parameter sequence, to generate an original geological parameter set; a space registration submodule, which performs registration on multiple parameters based on the original geological parameter set, calls time series and spatial coordinate information, calculates the spatial offset and time delay difference of parameters between multiple observation points, and performs coordinate alignment and interpolation correction on multiple source data, to generate a space-time registration parameter set; a parameter extraction submodule, which performs feature dimension reduction on rock density gradient, geothermal gradient, fracture extension rate, pore fluid pressure difference and magnetic field change according to the space-time registration parameter set, calculates the contribution rate of multiple parameters in the main component of mineralization, extracts the key mineralization feature sequence, and generates a geological parameter set.

4. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 1, characterized in that, The spatial strain module comprises: a gradient difference calculation submodule, which performs spatial distribution analysis based on the geological parameter set, extracts the coordinate position and parameter value sequence of adjacent sampling points, calculates the difference value of geological parameters between multiple points, performs standardization processing on the difference result according to the spatial distance weight, and generates a spatial parameter gradient difference value set; an intensity analysis submodule, which calls the spatial parameter gradient difference value set, calculates the directional vector amplitude of the geological parameter gradient in each region, analyzes the strain intensity value in the region according to the amplitude change interval, and analyzes in combination with the mean value change of the intensity of adjacent regions, to generate a spatial strain intensity distribution result; The tension point extraction submodule extracts a tension point set by identifying a region exceeding a spatial strain threshold based on the spatial strain intensity distribution result, and taking the central coordinates and corresponding intensity indicators as the tension point set, and calculates a spatial distance and a direction distribution of adjacent tension points to generate a spatial tension distribution result.

5. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 4, characterized in that, The spatial strain threshold is set based on a spatial parameter gradient difference and a corresponding strain intensity value calculated by obtaining a geological parameter change sequence in a plurality of typical regions, and a frequency distribution curve of the strain intensity.

6. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 1, characterized in that, The causal analysis module includes: The offset difference calculation submodule calculates an angle change amount between a plurality of parameters based on a direction vector of adjacent geological parameters and the spatial tension distribution result and the spatial coordinate sequence in the geological parameter set, judges a direction difference degree between a plurality of regions based on a direction offset reference value, generates a direction offset difference value sequence, and extracts a time sequence data in the direction offset difference value sequence and the geological parameter set to calculate a time sequence correlation coefficient in a same region, determines an abnormal interval of correlation based on a matching degree of the correlation coefficient and the direction offset difference, and obtains an abnormal correlation coefficient distribution result. The causal chain generation submodule filters a region in which a direction offset difference and a correlation direction are opposite based on the abnormal correlation coefficient distribution result, extracts a spatial position and a corresponding parameter offset feature, adjusts a causal direction order, rearranges a dependency relationship between a plurality of parameters, and generates a disturbance causal chain result. The direction offset reference value is set based on a mean value and a variance of an included angle calculated based on a change trend by calculating a direction vector between each pair of adjacent parameters in a plurality of regions.

7. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 6, characterized in that, The resonance response module includes:

8. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 1, characterized in that, The data analysis submodule obtains a plurality of geological parameter time sequence data in a disturbance region, detects a sampling interval and a data gap and completes a missing section, then rearranges according to a sampling time, calculates a change rate of a plurality of parameters between consecutive time points as a disturbance change rate, and generates a geological parameter time change sequence set. The phase synchronization submodule identifies an excitation event point based on the geological parameter time change sequence set, extracts an instantaneous phase value near the event point, calculates a phase difference between parameters and a stability degree in an event interval, analyzes an average synchronization relationship strength between parameters, and generates a phase synchronization rate distribution result. The resonance evaluation submodule extracts a parameter group with an optimal phase synchronization rate based on the phase synchronization rate distribution result, calculates a resonance intensity coefficient between a plurality of geological parameters, and inputs the resonance intensity coefficient and a disturbance amplitude value to weight and aggregate an influence strength of the parameter combination and sort the parameter combination to generate a preliminary prospecting result. The system further includes:

9. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 1, characterized in that, The optimization prospecting module performs joint analysis based on the preliminary prospecting result and the disturbance causal chain result, performs smoothing adjustment on a resonance coefficient distribution, eliminates isolated abnormal points through iterative updating, and calculates a mineralization factor score to generate an optimized prospecting analysis result. The optimized prospecting analysis result includes a smoothed resonance coefficient, a mineralization factor score, and isolated abnormal point elimination. The optimization prospecting module includes:

10. The artificial intelligence-based geological prospecting data processing and analysis system according to claim 9, characterized in that, ​ A resonance smoothing submodule extracts the resonance coefficient difference of adjacent intervals and performs amplitude constraint on the difference value based on the resonance coefficient of the same region in the preliminary prospecting result and the disturbance causal chain result, adjusts the resonance coefficient value of multiple intervals and reconstructs the distribution sequence, and generates a smooth resonance coefficient sequence; An iterative correction submodule calls the smooth resonance coefficient sequence, detects the distribution density of extreme values in the sequence and identifies isolated abnormal points, replaces the abnormal points with the mean value of adjacent intervals, recalculates the sequence change rate after correction and iterates multiple times until the density is stable, and obtains the resonance distribution result after correction of the anomaly; A metallogenic scoring submodule extracts three indexes of disturbance response amplitude, phase synchronization rate and resonance intensity coefficient according to the resonance distribution result after correction of the anomaly, calculates the weighted average of multiple indexes and synthesizes the metallogenic factor according to the weight ratio, analyzes and scores by region, and generates the optimized prospecting analysis result.

Images