Deep concealed deposit detection method based on multi-source geological data fusion

Abstract

The invention provides an innovative deep concealed ore deposit detection method, which comprehensively utilizes geology, geophysics, geochemistry and remote sensing technologies, and realizes effective integration of multi-source geological data by developing an advanced data fusion algorithm. According to the method, data collected by geological survey, wide-area electromagnetic method exploration, hydrocarbon-mercury superposition halo analysis and aviation hyperspectral remote sensing are subjected to standardized preprocessing, so that the consistency and accuracy of the data are ensured; and then, deep analysis is carried out on the fused data by utilizing machine learning and statistical analysis technologies so as to identify and predict mineralization modes and abnormal regions. Finally, a three-dimensional geologic model is constructed, the form and distribution of an ore deposit are visually displayed, and a scientific basis is provided for geological exploration. The method is characterized in that a multi-source data fusion technology and mineralization mode identification automation and precision are realized, so that the efficiency and the accuracy of deep hidden ore deposit detection are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of geological prospecting technology, in particular to a comprehensive detection method combining geology, geophysics, geochemistry and remote sensing technology, which is used to improve the detection accuracy and efficiency of deep concealed mineral deposits. Background Art

[0002] In the field of geological exploration, traditional methods of mineral deposit detection usually rely on single geological, geophysical data, geochemistry or remote sensing technology, which are often limited by the degree, depth and accuracy of detection work. For example, geological methods mainly analyze the distribution of rocks and minerals through surface observation and sampling, but this method is difficult to detect deep underground. Geophysical methods, such as seismic exploration or electromagnetic methods, can detect underground structures, but often require a lot of calculations and interpretation work, and have limited ability to identify deep hidden deposits. Geochemical methods analyze the chemical composition of soil and rocks to find signs of mineralization, but these methods are easily affected by changes in the surface environment and also face challenges in the detection of deep mineral deposits. Remote sensing technology, especially aerial or satellite remote sensing, can cover a large area of ​​the surface, but in identifying underground mineralization characteristics, traditional remote sensing technology has limited resolution and depth. In order to overcome these limitations and improve the efficiency and accuracy of the detection of deep hidden deposits, an innovative method that integrates multiple geological data is urgently needed. This method should be able to make full use of the advantages of geological, geophysical, geochemical and remote sensing data, and through advanced data processing and analysis technology, realize deeper and more accurate detection of underground mineralization characteristics under different mineralization geological conditions. It is in this context that the present invention is proposed, aiming to solve the shortcomings of existing technologies in the detection of deep concealed mineral deposits through technological innovation. Summary of the invention

[0003] The purpose of the present invention is to provide a comprehensive method for detecting deep concealed mineral deposits, which utilizes the advantages of geology, geophysics, geochemistry and remote sensing technology to achieve high-precision detection of underground mineralization characteristics through a series of innovative data processing and analysis processes.

[0004] The method for detecting deep concealed mineral deposits includes the following steps:

[0005] S1. Geological data collection: Use high-precision GPS equipment to record the latitude, longitude and elevation information of the sampling points. Through systematic field observation and sample collection, we record in detail key geological information such as rock type, mineral composition, stratigraphic structure, igneous rock, etc.

[0006] S2. Geophysical data acquisition: During the geophysical data acquisition phase, we use wide-area electromagnetic method technology to measure the resistivity and electromagnetic wave propagation characteristics of underground media through a ground coil system. The transmitting coil frequency range is set to 1Hz to 1kHz to detect underground structures at different depths. The resistivity measurement accuracy reaches 1%, ensuring the accurate acquisition of the electrical characteristics of the underground medium.

[0007] S3. Geochemical data collection: In the geochemical data collection step, we use high-precision analytical instruments to determine the types of hydrocarbon minerals, mercury content and other key elements in soil and rock samples. The analysis accuracy of mercury content in samples reaches the nanogram level (ng / g), and the content analysis of other elements such as gold (Au), copper (Cu), lead (Pb), zinc (Zn), etc. also reaches the corresponding high-precision standards. These data are crucial for identifying geochemical anomaly areas.

[0008] S4. Remote sensing data collection uses aerial hyperspectral remote sensing technology to conduct detailed spectral analysis of the target area, which can capture spectral data from visible light to near infrared or short-wave infrared, with a resolution usually at the nanometer level. These data enable us to identify and analyze the types of minerals in the area, including their chemical composition and physical properties. By analyzing the spectral characteristic curve, we can determine the unique spectral signature of a specific mineral, thereby identifying and analyzing the mineral, type, surface location, range and distribution characteristics.

[0009] S5. Data preprocessing: The data preprocessing step is to clean, standardize and denoise the collected multi-source geological data. The cleaning process includes removing invalid data and outliers, and the standardization process ensures that all data have a unified dimension and format. Denoising uses filtering algorithms such as wavelet transform to minimize data noise and improve data quality and consistency.

[0010] S6. Multi-source data fusion: In the multi-source data fusion step, we developed a fusion algorithm that performed data dimension reduction through principal component analysis (PCA), retained components with 95% variance explanation rate, and performed feature extraction through independent component analysis (ICA). In addition, multi-scale analysis technology was used to integrate data at different scales to comprehensively identify mineralization characteristics.

[0011] S7. Pattern recognition and analysis: In the pattern recognition and analysis phase, we used a variety of machine learning algorithms, including support vector machines (SVM), random forests, and deep learning networks. SVM uses a radial basis function (RBF) kernel and optimizes the classification effect by adjusting the parameters C and γ. Random forests construct a forest of 100 trees and use 10-fold cross validation to optimize the parameters. The deep learning network uses a convolutional neural network (CNN) structure with a network depth of 5 layers, and uses the ReLU activation function and Adam optimizer for training.

[0012] S8. Construction of 3D geological model: The 3D geological model construction step uses professional geological modeling software and combines geostatistical methods such as Kriging to construct a 3D geological model of the target area. The model takes into account geological structure, geophysical response and geochemical anomalies, optimizes spatial correlation parameters through simulated annealing algorithm, and predicts the spatial distribution characteristics of the deposit.

[0013] S9. Based on the potential deposit locations predicted by the model, we planned and executed precise drilling operations, collected core samples for lithology description and mineral composition analysis, and applied geophysical logging techniques, such as resistivity and natural gamma logging, to obtain the electrical characteristics of the underground medium. These measured data are used to compare carefully with the model's predictions, and any deviations will trigger further adjustments to the model. The adjustment process may involve re-evaluating the input parameters of the geological model, optimizing the parameters of the machine learning algorithm, such as the number of layers and neuron configuration of DBN, CNN, RNN, or adjusting the variance range and nugget value in the geostatistical method.

[0014] Preferably, the data preprocessing step first performs data calibration, and adjusts the raw data to eliminate measurement bias and systematic errors by comparing with known standards or reference samples to ensure that the calibration accuracy of the data reaches the expected standard. Next, outlier removal is performed, and statistical methods such as standard deviation analysis or box plots are used to identify outliers in the data set, which may be caused by measurement errors, sample contamination, or other atypical geological conditions. The outlier removal criteria are strictly set, for example, any value exceeding the mean ± 3 times the standard deviation will be regarded as an outlier and excluded. In addition, robust mathematical transformations such as logarithmic transformations or Z-score transformations are introduced to reduce the impact of extreme values ​​and improve the normality of the data.

[0015] Preferably, in the present invention, the optimization implementation of the data fusion algorithm is achieved through accurate principal component analysis (PCA) and independent component analysis (ICA) techniques. PCA is performed by calculating the covariance matrix Σ of the original data set X, and then performing eigenvalue decomposition to obtain eigenvectors vi and eigenvalues ​​λi, selecting the eigenvectors with the largest eigenvalues ​​of the first k, where k = 10, to form a transformation matrix W, thereby projecting the original data into a new space and retaining the principal components with a variance contribution rate of more than 85%. The mathematical expression is: W = [v1, v2, … ,vk]Xnew=XW In the ICA stage, the algorithm iteratively maximizes the independence of independent source signals of non-Gaussian distribution, uses Curtosis or Negentropy as the independence measure, and sets the convergence standard as the change of independence index less than ∈=0.001. The iterative update formula of ICA can be expressed as: Where w is the weight vector, s is the signal extracted from the data after PCA, and t is the number of iterations. This process continues until the convergence condition is met or the maximum number of iterations is reached, which is usually no more than 1000. Through this method of combining PCA and ICA, the data fusion algorithm not only improves the interpretability of the data, but also significantly enhances the ability to identify key features of deep concealed deposits through precise mathematical models and strict parameter control. This method ensures that the most informative features are extracted from multi-source geological data, providing a solid foundation for subsequent mineral deposit exploration and resource evaluation.

[0016] Preferably, in the pattern recognition and analysis step, a series of advanced machine learning models are used to improve the recognition accuracy of mineralization features. These models include deep belief networks (DBN), convolutional neural networks (CNN) and recurrent neural networks (RNN), each of which is optimized for different characteristics of the data and analysis requirements. DBN is stacked with multiple layers of restricted Boltzmann machines (RBM), each layer containing 1,000 neurons, pre-training features through unsupervised learning, and then fine-tuning through supervised learning to automatically extract complex patterns of data. The CNN model contains 5 convolutional layers and 3 fully connected layers, using ReLU activation function and cross entropy loss function, and the convolution kernel size is set to 3x3 with a step size of 1 to capture local spatial features and perform effective feature mapping. RNN, especially long short-term memory network (LSTM), contains 3 layers, 128 neurons in each layer, which is used to process sequence data and capture long-term dependencies in time series. All models use the Adam optimizer for parameter update, the learning rate is set to 0.001, overfitting is controlled by early stopping, and training is stopped when the performance improvement on the validation set is less than 0.01. The integrated use of these models enables us to analyze geological data from multiple angles and levels, significantly improving the accuracy and efficiency of mineral deposit detection.

[0017] Preferably, in the three-dimensional geological model construction step, a geostatistical method is used to enhance the prediction accuracy of the model. Specifically, the Kriging method is used for interpolation and prediction of spatial data, and the correlation of spatial variables is described by establishing a semivariance function, wherein we selected a spherical model as the semivariance function, and its parameters include a range set to 500 meters, a sill set to 1.5 times the coefficient of variation, and a nugget effect set to 10% of the nugget to ensure the sensitivity of the model to close spatial correlation. In addition, a simulated annealing algorithm is applied to the parameter optimization of the geostatistical model to avoid local optimal solutions, with an initial temperature set to 1000, a cooling rate set to 0.98, and the Metropolis criterion in the acceptance criterion used to determine the acceptance probability of the new solution. Through the fine adjustment of these parameters and the application of geostatistical methods, the three-dimensional geological model can more accurately reflect the spatial distribution characteristics of underground mineralization, thereby significantly improving the prediction accuracy and reliability of ore deposit detection.

[0018] Preferably, in the verification and optimization step, comprehensive exploration data is used for field verification and optimization of the model. Specifically, this step covers the analysis of drilling cores, including detailed records of the lithology description, mineral composition, structure and mineralization characteristics of the cores. The drilling depth reaches 1,500 meters, and the core sampling rate exceeds 90%, ensuring the accurate acquisition of underground geological information. The collection of geophysical logging data uses a variety of logging methods such as resistivity, natural gamma, density and neutron porosity, and the measurement accuracy is controlled within 1% to verify the electrical characteristics of the geological model. The geochemical sampling results analyze the element content in soil, water and rock samples, such as mercury, copper, lead, zinc, etc., with an accuracy of ppb (parts per billion) level to identify geochemical anomalies. The comprehensive analysis of these exploration data provides actual geological, geophysical and geochemical evidence for the geological model, enabling us to accurately adjust and optimize the model, thereby improving the accuracy and reliability of mineral deposit detection.

[0019] The deep concealed mineral deposit detection method of the present invention proposes a highly integrated detection system, which has carefully designed multiple functional modules to achieve full-process automation and high efficiency. The system includes a data acquisition module, which is responsible for the on-site collection of geological, geophysical, geochemical and remote sensing data; a data preprocessing module, which ensures data quality, performs cleaning, standardization and outlier removal, and the data calibration accuracy is controlled within 0.5%; a data fusion module, which applies PCA and ICA techniques for data dimension reduction and feature extraction, PCA retains components with more than 90% variance, and the ICA independence index change threshold is set to 0.0001; a pattern recognition module, which integrates machine learning algorithms such as DBN, CNN and RNN, uses the Adam optimizer, and the learning rate is set to 0.001; a three-dimensional geological modeling module, which uses the Kriging method for spatial data interpolation, and the simulated annealing algorithm optimizes parameters, with the initial temperature set to 1000 and the cooling factor set to 0.98; a verification and optimization module, which combines 1500-meter-deep drilling core analysis, 1% precision geophysical logging data and ppb-level precision geochemical sampling results to conduct field verification and optimization of the model. The collaborative work of these modules ensures the automation of the entire process from data collection to final model verification, significantly improving the efficiency and accuracy of mineral deposit detection.

[0020] According to the deep concealed mineral deposit detection method described in claims 1 to 7, the computer program product of the present invention is an integrated package containing professional software codes, which are designed to automatically perform key steps such as data acquisition, preprocessing, fusion, pattern recognition, three-dimensional geological modeling, and verification and optimization. The software is developed using an efficient programming language to ensure fast processing and high compatibility, and follows industry coding standards. The product is stored in the form of a CD, hard disk or solid-state drive, and users can deploy and use it on a personal computer or server through simple installation steps. The software has a user-friendly interface and provides a detailed user manual and online help document to support users to quickly master the operation method. The program adopts a modular design to facilitate future functional expansion and updating, ensuring the long-term effectiveness and adaptability of the software. This computer program product provides users with a comprehensive, accurate and easy-to-use solution for deep concealed mineral deposit detection, significantly improving the efficiency and accuracy of geological exploration work.

[0021] Further, the method or system for detecting deep concealed mineral deposits, the present invention also includes an advanced user interface, which provides users with an intuitive and easy-to-operate interactive platform. The interface allows users to input geological parameters, such as rock type, mineral composition, stratigraphic age, etc., and select analysis methods such as data preprocessing options, fusion algorithms, and machine learning models. Users can view the data processing flow, pattern recognition results, and the construction process and final results of the three-dimensional geological model in real time, all of which are displayed graphically to ensure the clarity and comprehensibility of the information. The user interface also has a model adjustment function, providing graphical adjustment tools such as sliders, drop-down menus, and buttons, so that users can fine-tune model parameters directly in the interface without complex programming or command input. The adjusted results can be displayed immediately on the interface, which is convenient for users to quickly evaluate the adjustment effect. The interface also integrates logging and history recording functions to record all operation steps and model status, and supports users to backtrack and compare versions. In order to ensure the sharing and further analysis of the results, the user interface supports multiple data output formats, including PDF, CSV, and DXF, so that users can easily export model results and related data. Overall, this user interface design greatly improves the usability and flexibility of the system, making the detection of deep concealed mineral deposits more efficient and accurate, while providing users with a rich interactive experience.

[0022] The method or system for detecting deep concealed mineral deposits, the present invention also includes an innovative geological exploration decision support tool. The tool is designed to prioritize potential mineral deposits and make intelligent recommendations for resource allocation based on the prediction results obtained by the aforementioned method. It uses advanced algorithms to evaluate the mineralization potential and exploration value of each region by analyzing three-dimensional geological models, geochemical anomalies, geophysical responses and other relevant geological data. The decision support tool adopts a multi-standard decision analysis method, comprehensively considers factors such as geological characteristics, exploration costs, expected returns and risk levels, and automatically generates an exploration priority list. It can also provide customized resource allocation plans based on the size of the exploration team, equipment configuration and project schedule to ensure efficient organization and execution of exploration activities. The tool has a user-friendly interface that allows the geological exploration team to easily input relevant parameters, adjust evaluation criteria and optimize recommended results. It also supports integration with existing exploration management software and databases to achieve seamless docking and real-time updating of data. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 The data preprocessing flow chart of the deep concealed mineral deposit detection method of the present invention is

[0024] Figure 2 The step diagram of the deep hidden ore deposit detection method of the present invention is as follows DETAILED DESCRIPTION

[0025] The purpose of the present invention is to provide a comprehensive method for detecting deep concealed mineral deposits, which utilizes the advantages of geology, geophysics, geochemistry and remote sensing technology to achieve high-precision detection of underground mineralization characteristics through a series of innovative data processing and analysis processes.

[0026] The method for detecting deep concealed mineral deposits includes the following steps:

[0027] S1. Geological data collection: High-precision GPS equipment is used to record the latitude, longitude and elevation information of the sampling points. Through systematic field observation and sample collection, we record in detail key geological information such as rock type, mineral composition, stratigraphic structure, igneous rock, etc. Geological samples are subjected to laboratory analysis, such as X-ray diffraction (XRD), to determine the quantitative composition of minerals, providing basic geological information for subsequent data processing and deposit prediction.

[0028] S2. Geophysical data acquisition: During the geophysical data acquisition phase, we use wide-area electromagnetic method technology to measure the resistivity and electromagnetic wave propagation characteristics of underground media through a ground coil system. The transmitting coil frequency range is set to 1Hz to 1kHz to detect underground structures at different depths. The resistivity measurement accuracy reaches 1%, ensuring the accurate acquisition of the electrical characteristics of the underground medium.

[0029] S3. Geochemical data collection: In the geochemical data collection step, we use high-precision analytical instruments to measure the mercury content and other key elements in soil, water and rock samples. The mercury content analysis accuracy in soil samples reaches the nanogram level (ng / g), and the content analysis of other elements such as gold (Au), copper (Cu), lead (Pb), zinc (Zn), etc. also reaches the corresponding high-precision standards. These data are crucial for identifying geochemical anomaly areas.

[0030] S4. Remote sensing data collection: During the remote sensing data collection phase, we use aerial hyperspectral remote sensing technology to collect spectral feature data of the target area. The sensor used has a high spatial resolution of better than 5 meters and a wide wavelength range of 400nm to 2500nm, which provides rich information for analyzing the surface mineral composition and surface coverage, thereby identifying and analyzing minerals, types, surface locations, ranges and distribution characteristics.

[0031] S5. Data preprocessing: The data preprocessing step is to clean, standardize and denoise the collected multi-source geological data. The cleaning process includes removing invalid data and outliers, and the standardization process ensures that all data have a unified dimension and format. Denoising uses filtering algorithms such as wavelet transform to minimize data noise and improve data quality and consistency.

[0032] S6. Multi-source data fusion: In the multi-source data fusion step, we developed a fusion algorithm that performed data dimension reduction through principal component analysis (PCA), retained components with 95% variance explanation rate, and performed feature extraction through independent component analysis (ICA). In addition, multi-scale analysis technology was used to integrate data at different scales to comprehensively identify mineralization characteristics.

[0033] S7. Pattern recognition and analysis: In the pattern recognition and analysis phase, we used a variety of machine learning algorithms, including support vector machines (SVM), random forests, and deep learning networks. SVM uses a radial basis function (RBF) kernel and optimizes the classification effect by adjusting the parameters C and γ. Random forests construct a forest of 100 trees and use 10-fold cross validation to optimize the parameters. The deep learning network uses a convolutional neural network (CNN) structure with a network depth of 5 layers, and uses the ReLU activation function and Adam optimizer for training.

[0034] S8. Construction of 3D geological model: The 3D geological model construction step uses professional geological modeling software and combines geostatistical methods such as Kriging to construct a 3D geological model of the target area. The model takes into account geological structure, geophysical response and geochemical anomalies, optimizes spatial correlation parameters through simulated annealing algorithm, and predicts the spatial distribution characteristics of the deposit.

[0035] S9. In the verification and optimization step, we accurately verify the location of the ore deposit predicted by the model through field drilling and geophysical logging. The specific operation includes drilling in the mineralized area predicted by the model. The drilling depth is determined according to the geological model prediction and geological structure analysis, usually reaching 1,500 meters, and the core sampling rate exceeds 95%. During the drilling process, a detailed lithology description and mineral composition analysis are carried out every 10 meters, and the accurate identification of mineral composition is ensured by using technologies such as X-ray diffraction (XRD) and electron probe microanalysis (EPMA). Geophysical logging data, including resistivity and natural gamma logging, are controlled within 1% to obtain the electrical characteristics of the underground medium. The collected drilling and geophysical data are compared and analyzed with the model prediction results. Any deviation between the prediction and the actual data will trigger the adjustment of the model. For example, if the mineralized area predicted by the model is not found in the drilling, we will re-evaluate the input parameters of the geological model and adjust the parameters in the machine learning algorithm, such as increasing the number of layers from 3 to 5 in DBN, adjusting the convolution kernel size from 3x3 to 5x5 in CNN, and adjusting the step size from 1 to 2, or increasing the time step from 5 to 10 in RNN and the number of neurons from 128 to 256. Kriging parameters in geostatistical methods, such as the variance range from 200 meters to 500 meters and the nugget value from 1.5 times the coefficient of variation of the data to 2 times, will also be optimized based on actual data. Through this iterative verification and optimization process based on parameter values, we can significantly improve the prediction accuracy and reliability of the ore deposit detection model and ensure the effective application of the model in actual geological exploration.

[0036] Data preprocessing First, perform data calibration, and adjust the original measurement values ​​to within the standard error range by comparing with international geological sample standards, usually not exceeding ±0.5% relative error; second, perform outlier removal, and determine 1.5 times of IQR (interquartile range) as the threshold by calculating the quartiles and interquartile range of the data set. Any data point exceeding the 25th percentile minus 1.5 times IQR or exceeding the 75th percentile plus 1.5 times IQR is considered an outlier and excluded; in addition, the Z score method is applied to further identify and remove outliers, and points with an absolute value of Z score greater than 3 are considered outliers. Through these precise parameter settings and strict quality control measures, the data preprocessing steps ensure the accuracy and reliability of the data set, providing a solid foundation for subsequent multi-source data fusion and pattern recognition.

[0037] The data fusion algorithm achieves data dimension reduction and feature selection through precise principal component analysis (PCA) and independent component analysis (ICA) techniques. Specifically, in the PCA stage, the algorithm first calculates the eigenvalues ​​and eigenvectors of the covariance matrix, then sorts them according to the size of the eigenvalues ​​and selects the first five principal components, whose cumulative contribution rate exceeds 90%, ensuring that the main information of the data set is retained. In the ICA stage, the algorithm initializes the weight matrix and optimizes iteratively until the change in independence indicators (such as mutual information) is less than 0.0001, ensuring the stability and reliability of the independent components. The number of iterations of ICA usually does not exceed 1,000 to avoid overfitting. Through this method of combining PCA and ICA, the data fusion algorithm not only significantly reduces the dimensionality of the data, but also accurately extracts features that are critical to mineral deposit detection, providing high-quality input data for subsequent pattern recognition and geological model construction.

[0038] Deep hidden mineral deposit detection method, in the pattern recognition and analysis step, we carefully designed and implemented a series of machine learning models to achieve efficient recognition of mineralization features in geological data. The deep belief network (DBN) consists of 5 layers of stacked restricted Boltzmann machines (RBM), each layer contains 500 neurons, pre-trained by contrastive divergence algorithm, the learning rate is set to 0.01, and then fine-tuned to optimize the extraction of mineralization features. The convolutional neural network (CNN) consists of 3 convolutional layers and 2 fully connected layers. The convolutional layer uses a 3x3 convolution kernel with a step size of 1. Each convolutional layer is followed by a maximum pooling layer to reduce the feature dimension and enhance the invariance of the feature. The ReLU activation function is used to introduce nonlinearity, and the learning rate is set to 0.001. Recurrent neural network (RNN), specifically long short-term memory network (LSTM), contains 3 layers, each layer has 128 neurons, to process geological time series data and capture the long-term dependencies of geological events. The time step is set to 5, and the learning rate of forget gate and input gate is set to 0.0005. All models use cross entropy as the loss function and are trained by Adam optimizer. The optimizer's β1 and β2 parameters are set to 0.9 and 0.999 respectively to adaptively adjust the learning rate.

[0039] In the three-dimensional geological model construction step, we used geostatistical methods, including Kriging and simulated annealing algorithms, to significantly improve the prediction accuracy of the model. In the Kriging method, we set the parameters of the semivariance function, such as the range parameter of 500 meters, the sill parameter of 1.5 times the coefficient of variation of the data, and the baseline effect parameter of 5% of the nugget, to accurately capture the correlation of spatial data. The parameter setting of the simulated annealing algorithm includes an initial temperature of 1000 degrees, followed by a cooling factor of 0.98 to gradually reduce the temperature, ensuring that the search process traverses enough solution space to avoid falling into the local optimum. The probability of accepting a new solution is determined by the Metropolis criterion, in which the temperature parameter and the energy difference jointly determine the acceptance probability.

[0040] In the verification and optimization steps, we took a series of specific exploration data collection and analysis measures. First, the drilling core analysis depth was set at 1,500 meters, and the core sampling rate target was 95%, ensuring the comprehensiveness and continuity of underground geological information. During the drilling process, detailed core description and sampling were carried out every 10 meters, including the recording of rock type, mineral composition, structure and mineralization characteristics. Secondly, geophysical logging data collection covers parameters such as resistivity, natural gamma, density and neutron porosity, and the measurement accuracy of each parameter is controlled within 1% to ensure high-precision comparison with the electrical characteristics of the geological model. In addition, the geochemical sampling results are analyzed by analyzing the content of key elements such as mercury, copper, lead, zinc, etc. in soil, water and rock samples, with an accuracy of ppb level, such as 1ppb for mercury content analysis and 5ppb for copper content analysis, to identify geochemical anomalies. Through these detailed exploration data and precise analysis, we are able to strictly verify and optimize the geological model in a targeted manner, thereby significantly improving the accuracy and reliability of mineral deposit detection.

[0041] The system implemented by the present invention has a series of precisely set modules to ensure the efficiency and accuracy of the entire detection process. The system includes a data acquisition module for collecting geological, geophysical, geochemical and remote sensing data; a data preprocessing module for data cleaning, standardization and outlier removal, with data calibration accuracy controlled within ±0.5%, and the outlier removal standard set to exceed the mean ±3 times the standard deviation; a data fusion module that uses PCA to retain components with more than 90% of the explained variance, and the ICA independence index change threshold is set to 0.0001; a pattern recognition module that uses DBN, CNN and RNN algorithms, where CNN contains 3 layers of convolutional neural networks. The LSTM network has 128 neurons per layer, and all models use the Adam optimizer with a learning rate of 0.001. In the 3D geological modeling module, the range parameter of the Kriging method is set to 500 meters, the nugget parameter is 1.5 times the coefficient of variation of the data, the initial temperature of the simulated annealing algorithm is set to 1000, and the cooling factor is 0.98. In the verification and optimization module, combined with the analysis of 1500-meter-deep drilling cores, the accuracy of geophysical logging data is controlled within 1%, and the accuracy of geochemical analysis reaches 1ppb. The close integration of these modules and the precise setting of parameters have realized the automation and optimization of the entire process from data collection to model verification, greatly improving the efficiency and accuracy of mineral deposit detection.

[0042] The computer program product of the present invention is a highly specialized software package containing precisely written codes that can automatically perform a series of complex operations including data acquisition, preprocessing, data fusion, pattern recognition, three-dimensional geological modeling, verification and optimization. The software code adopts a modular structure and is implemented in efficient programming languages ​​such as C++ or Python, ensuring the execution efficiency and cross-platform compatibility of the algorithm. The program product provides a user interface and supports display at least 1024x768 resolution, ensuring that users can obtain clear visual effects on different devices. The software product is stored in a computer-readable medium, such as a DVD disc (capacity not less than 4.7GB), a hard disk (rotation speed not less than 7200RPM) or a solid-state drive (write speed not less than 500MB / s) to meet the storage and reading needs of different users. During the installation process, the software provides a detailed installation wizard, supports at least Windows 10 and Linux Ubuntu 20.04 operating systems, and the memory requirement is not less than 4GB RAM, ensuring the stable operation of the software. The software package also includes a user manual with a word count of not less than 200,000 words, which details the installation, configuration, operation process and troubleshooting guide of the software. The online help documentation provides at least 1,000 answers to frequently asked questions, ensuring that users can quickly resolve problems encountered during use.

[0043] The system of the present invention is equipped with a powerful user interface, which allows users to input geological parameters, select analysis methods, and view and adjust the results of the three-dimensional geological model in real time through a concise and clear graphical user interface (GUI). The user interface provides parameter input fields, including but not limited to rock type selectors, mineral composition input boxes (supporting input of up to 100 minerals), stratigraphic age timestamps (accurate to millions of years), and detailed description areas for geological structure data. Users can select data preprocessing algorithms through a drop-down menu. For example, when selecting principal component analysis (PCA), the number of components to be retained (such as the first 5 principal components) can be selected, or the independence index change threshold (such as 0.0001) can be set in independent component analysis (ICA). In the pattern recognition module, users can select different machine learning models, such as deep belief networks (DBNs), convolutional neural networks (CNNs), and set the number of network layers and the number of neurons. The number of convolutional layers in CNN can be set to 3 to 5 layers, and the convolution kernel size of each layer can be set to 3x3 or 5x5. The user interface also includes a real-time results display area that can display a rendered image of the 3D geological model, supporting a display of at least 1024x768 resolution, ensuring that model details are clearly visible. The model adjustment tool allows users to adjust the parameters of the kriging method using sliders, such as the range can be adjusted from 100 meters to 1000 meters, and the nugget and baseline effect parameters can also be adjusted using percentage sliders. The user interface also has a detailed logging function that records the user's every step of operation and model status, and supports historical record back and version comparison. The data output function supports exporting results to PDF reports, CSV data tables, and DXF drawing formats, ensuring that users can easily share and further analyze the results.

[0044] The present invention provides a geological exploration decision support tool, which uses a quantitative decision-making algorithm based on the prediction results to provide priority sorting and resource allocation suggestions for mineral deposit exploration. This tool comprehensively considers the multi-dimensional characteristics of geological data, such as the probability value of the geological model (for example, the mineralization probability threshold is set to more than 70%), the intensity of geochemical anomalies (such as mercury content exceeding 1ppb), the resistivity anomaly of geophysical measurements (set to 20% lower than the surrounding background value), and other parameters, as well as exploration costs and expected economic benefits, and automatically evaluates the exploration value of each potential mineral deposit through a weighted scoring system.

[0045] The decision support tool uses Monte Carlo simulation to assess risk and uncertainty, with the number of simulations set to 1,000 to ensure the reliability of statistical results. It also combines multi-objective optimization algorithms such as genetic algorithms, with the number of iterations set to 100 generations and the crossover rate and mutation rate set to 0.8 and 0.01, respectively, to find the optimal resource allocation solution.

[0046] The user interface allows the exploration team to adjust parameters according to actual needs, such as adjusting the weight of exploration costs or the discount rate of expected returns, to ensure that the recommended solution is in line with the actual exploration strategy. The tool also provides a real-time feedback mechanism to dynamically update the recommended results based on exploration progress and market changes.

[0047] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for detecting deep concealed mineral deposits, characterized in that: The method comprises the following steps: S1. Geological data acquisition: Through data collection, field investigation and sampling, collect and obtain key geological information such as rock type, mineral composition, stratigraphic structure, igneous rock, etc. in the target area. S2. Geophysical data acquisition: Use wide-area electromagnetic technology to measure geophysical parameters such as the resistivity of underground media and electromagnetic wave propagation characteristics to invert deep geological characteristics. S3. Geochemical data collection: hydrocarbon-mercury superposition halo analysis technology is used to collect hydrocarbon mineral combinations, mercury content and other geochemical indicators in soil and rocks. S4. Remote sensing data acquisition: Use aerial hyperspectral remote sensing technology to obtain spectral characteristic data of the target area to identify and analyze minerals, types, surface locations, ranges and distribution characteristics. S5. Data preprocessing: Clean, standardize and denoise the collected geological, geophysical, geochemical and remote sensing data to improve data quality and consistency. S6. Multi-source data fusion: Develop a fusion algorithm to perform feature extraction and multi-scale analysis on the pre-processed data and integrate them into a unified data set. S7. Pattern recognition and analysis: Use machine learning algorithms, such as support vector machines, random forests, or deep learning networks, to perform pattern recognition on the fused data and analyze mineralization characteristics and abnormal areas. S8. Construction of three-dimensional geological model: Based on the identification results of mineralization patterns and abnormal areas, a three-dimensional geological model of the target area is constructed using geological modeling software to predict the spatial distribution of mineral deposits. S9. Verification and optimization: The location of the mineral deposits predicted by the model is verified on the spot through drilling or geophysical verification, and the model is adjusted and optimized based on the verification results.

2. The method according to claim 1, characterized in that The data preprocessing step further includes data calibration and outlier removal to ensure the accuracy and reliability of the data.

3. The method according to claim 1, characterized in that The data fusion algorithm specifically includes principal component analysis (PCA) and independent component analysis (ICA) techniques, which are used for data dimension reduction and feature selection.

4. The method according to claim 1, characterized in that: In the pattern recognition and analysis step, the machine learning model includes but is not limited to a deep belief network (DBN), a convolutional neural network (CNN) or a recurrent neural network (RNN).

5. The method according to claim 1, characterized in that The three-dimensional geological model building step uses geostatistical methods, such as Kriging or simulated annealing algorithm, to improve the prediction accuracy of the model.

6. The method according to claim 1, characterized in that In the verification and optimization steps, the actual exploration data include drilling core analysis, geophysical logging data and geochemical sampling results.

7. A system for implementing the method according to any one of claims 1 to 6, characterized in that: The system includes but is not limited to the following modules: a data acquisition module, a data preprocessing module, a data fusion module, a pattern recognition module, a three-dimensional geological modeling module and a verification and optimization module.

8. A computer program product, characterized in that Comprising software code for executing the method according to any one of claims 1 to 7, said code being stored on a computer readable medium, such as an optical disk, a hard disk or a solid state drive.

9. The method or system according to any one of claims 1 to 8, characterized in that: The system also includes a user interface that allows the user to input geological parameters, select analysis methods, view model results, and make model adjustments.

10. A decision support tool for geological exploration, characterized in that: The tool provides mineral deposit exploration priority sorting and resource allocation suggestions based on the prediction results of the method described in any one of claims 1 to 9.

Images