A method for identifying magnetic susceptibility for deep structure analysis of a mining area

Abstract

The application discloses a magnetic susceptibility identification method for deep structure analysis of a mining area, and relates to the technical field of geological mineral exploration. The magnetic susceptibility identification method for deep structure analysis of a mining area comprises the following steps: data acquisition and preprocessing, model construction and inversion, magnetic susceptibility anomaly identification, and structure analysis and classification. According to the initial three-dimensional geological model and through the layered mixed inversion method, the magnetic susceptibility is inverted, the advantages of different inversion algorithms can be fully exerted, the shortcomings of a single algorithm can be made up, the accuracy of the magnetic susceptibility inversion result is improved, the efficient local algorithm and the strong global algorithm can be respectively used according to the geological characteristics of the shallow area and the deep area, the calculation efficiency and stability of the magnetic susceptibility inversion are improved, the set buffer area can smoothly transition and fuse the results of different magnetic susceptibility inversions, the continuous and reliable inversion result is ensured, the subsequent magnetic susceptibility anomaly identification is facilitated, and the identification effect of the magnetic susceptibility identification method is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geological mineral exploration, and particularly relates to a magnetic susceptibility identification method for deep structure analysis of a mining area. BACKGROUND

[0002] Magnetic susceptibility is a physical quantity that measures the degree of magnetization of a material in an applied magnetic field, reflecting the magnetic properties of the material. In deep structure analysis of a mining area, the magnetic susceptibility identification method helps to identify the magnetic characteristics of different rock layers and geological structures underground by measuring and analyzing the magnetic susceptibility of rocks. The principle is based on the strong correlation between rock magnetic susceptibility and geological structure. Through multi-dimensional, high-precision magnetic susceptibility data collection and intelligent analysis, the spatial distribution, morphology and physical characteristics of the deep structure of the mining area are revealed. Common methods include magnetic susceptibility sounding and magnetic susceptibility inversion technology, which effectively detect ore bodies and structures in combination with geophysical data and geological models, and provide key geological information for mining exploration and development.

[0003] The existing method for analyzing the structure of the deep part of the mining area based on the magnetic susceptibility inversion technology usually first uses the conjugate gradient method to invert the magnetic susceptibility distribution of the deep part of the mining area, and then uses threshold segmentation and other methods to identify the magnetic susceptibility anomaly area of the deep part of the mining area, and infers the geological structure type of the magnetic susceptibility anomaly area based on the geological structure characteristics. However, when the conjugate gradient method and other local optimization algorithms are used to invert the magnetic susceptibility of the mining area, distortion may occur due to complex terrain, thereby affecting the accuracy of the magnetic susceptibility inversion. In addition, the threshold segmentation and other methods for identifying anomalies in magnetic susceptibility are also susceptible to data distribution and local noise interference, which may result in the omission of small-scale anomalies, thereby affecting the identification effect of the magnetic susceptibility identification method for deep structure analysis of the mining area.

[0004] Based on the above, the present application provides a magnetic susceptibility identification method for deep structure analysis of a mining area with good identification effect. SUMMARY

[0005] In order to overcome the shortcomings of the existing method for analyzing the structure of the deep part of the mining area based on the magnetic susceptibility inversion technology, which uses the conjugate gradient method and other local optimization algorithms to invert the magnetic susceptibility of the mining area, and is susceptible to distortion due to complex terrain, thereby affecting the accuracy of the magnetic susceptibility inversion, and is also susceptible to data distribution and local noise interference when using threshold segmentation and other methods to identify anomalies in magnetic susceptibility, which may result in the omission of small-scale anomalies, thereby affecting the identification effect of the magnetic susceptibility identification method for deep structure analysis of the mining area, the present application provides a magnetic susceptibility identification method for deep structure analysis of a mining area with good identification effect.

[0006] The magnetic susceptibility identification method for deep structure analysis of a mining area comprises the following steps:

[0007] Data acquisition and preprocessing, collecting geological data of a mine area to be analyzed and obtaining magnetic susceptibility measurement data from the surface and drill holes of the mine area through a magnetic susceptibility sensor, preprocessing the magnetic susceptibility measurement data to obtain preprocessed magnetic susceptibility measurement data;

[0008] Model construction and inversion, constructing an initial three-dimensional geological model based on the geological data of the mine area to be analyzed and the preprocessed magnetic susceptibility measurement data, and inverting the magnetic susceptibility according to the initial three-dimensional geological model through a hierarchical mixed inversion method to obtain magnetic susceptibility distribution data of the mine area to be analyzed;

[0009] Magnetic susceptibility anomaly identification, identifying abnormal areas of the magnetic susceptibility distribution of the mine area to be analyzed based on the magnetic susceptibility distribution data through a multi-scale fusion hierarchical segmentation to obtain an identification result of the magnetic susceptibility abnormal area, wherein the identification result includes the position, shape and size of the magnetic susceptibility abnormal area;

[0010] Structural analysis and classification, extracting features of the identified magnetic susceptibility abnormal area to obtain structural features of the magnetic susceptibility abnormal area, and inputting the structural features of the magnetic susceptibility abnormal area into a trained geological structure classification model to obtain a geological structure classification result of the magnetic susceptibility abnormal area.

[0011] As a preferred aspect of the invention, the specific steps of preprocessing the magnetic susceptibility measurement data to obtain preprocessed magnetic susceptibility measurement data are:

[0012] Wavelet transform denoising processing, wavelet decomposition of the magnetic susceptibility measurement data to obtain high-frequency and low-frequency parts of the magnetic susceptibility measurement data, retaining the low-frequency part containing geological information and removing the high-frequency part containing noise, and reconstructing the denoised magnetic susceptibility measurement data through inverse discrete wavelet transform;

[0013] Inverse distance weighted spatial interpolation, obtaining position information of each data point in the magnetic susceptibility measurement data, determining a target position that needs to be interpolated, calculating the Euclidean distance between the target position and all known data points, selecting a power parameter and setting the distance weight of all known data points based on the Euclidean distance, and calculating the magnetic susceptibility value of the target position according to the distance weight;

[0014] Moving average smoothing processing, selecting a fixed length sliding window, starting from the beginning of the magnetic susceptibility measurement data sequence, moving one data point at a time, calculating the arithmetic mean value of all data points in the current sliding window as the smoothed value corresponding to the center position of the sliding window, and arranging the average values calculated by each sliding window in sequence to form a new smoothed data sequence.

[0015] As a preferred aspect of the application, the specific steps of constructing the initial three-dimensional geological model based on the geological data of the mine area to be analyzed and the preprocessed magnetic susceptibility measurement data are as follows:

[0016] Collecting geological data, obtaining geological data of the mine area to be analyzed, including drilling data, geological profile and surface geological map of the mine area to be analyzed;

[0017] Dividing stratum units, dividing the main stratum units of the mine area to be analyzed and determining the magnetic susceptibility range of each unit according to the collected geological data of the mine area to be analyzed;

[0018] Drawing structural interfaces, drawing the main structural interfaces of the mine area to be analyzed according to the geological profile in the geological data of the mine area to be analyzed, including faults and rock layer contact surfaces;

[0019] Three-dimensional geological modeling, based on the above-obtained geological structure information of the mine area and the preprocessed magnetic susceptibility measurement data and through a three-dimensional modeling software, an initial three-dimensional geological model is constructed.

[0020] As a preferred aspect of the application, the specific steps of inverting the magnetic susceptibility according to the initial three-dimensional geological model and through the hierarchical hybrid inversion method to obtain the magnetic susceptibility distribution data of the mine area to be analyzed are as follows:

[0021] Using finite difference or finite element numerical method and according to the initial three-dimensional geological model, the theoretical magnetic susceptibility data of each magnetic susceptibility measurement point is calculated, the theoretical magnetic susceptibility data calculated by forward calculation is compared with the actual preprocessed magnetic susceptibility measurement data, and the difference between the two is analyzed;

[0022] According to the geological characteristics and depth, the mine area to be analyzed is divided into layers, a target function is constructed in combination with regularization constraints, local optimization algorithm and global optimization algorithm are selected, and magnetic susceptibility inversion is performed on the layered mine area to be analyzed through the hierarchical hybrid inversion method, the parameters of the initial three-dimensional geological model are adjusted through iterative optimization, and the target function is minimized, wherein in each iterative optimization process, the parameters of the new three-dimensional geological model are calculated by forward calculation, until the error between the theoretical magnetic susceptibility data and the actual preprocessed magnetic susceptibility measurement data meets the convergence condition, and the magnetic susceptibility distribution data of the mine area to be analyzed is obtained.

[0023] As a preferred aspect of the application, the specific steps of dividing the mine area to be analyzed according to the geological characteristics and depth, constructing the target function in combination with regularization constraints, selecting the local optimization algorithm and the global optimization algorithm, and performing the magnetic susceptibility inversion on the layered mine area to be analyzed through the hierarchical hybrid inversion method are as follows:

[0024] According to the geological characteristics and depth, the mine area to be analyzed is divided into shallow, buffer and deep regions, and a target function is constructed in combination with regularization constraints;

[0025] The local optimization algorithm is selected to perform the magnetic susceptibility inversion on the shallow area of the mine area to be analyzed, and the global optimization algorithm is selected to perform the magnetic susceptibility inversion on the deep area of the mine area to be analyzed.

[0026] For the buffer area of the mine area to be analyzed, the local optimization algorithm and the global optimization algorithm are used to perform the magnetic susceptibility inversion on the buffer area respectively, to obtain the magnetic susceptibility inversion results of the two algorithms, the weights of the magnetic susceptibility inversion results of the two algorithms are set according to the geological depth, and the two results are weighted and fused to obtain the magnetic susceptibility inversion result of the buffer area, wherein the weight of the local optimization algorithm is the proportion of the distance between the data point to be inverted and the deep area to the overall width of the buffer area, and the weight of the global optimization algorithm is the proportion of the distance between the data point to be inverted and the shallow area to the overall width of the buffer area.

[0027] As a preferred aspect of the application, the specific steps of identifying the abnormal area of the magnetic susceptibility distribution of the mine area to be analyzed based on the magnetic susceptibility distribution data and through the hierarchical segmentation of multi-scale fusion are as follows:

[0028] The decomposition level of the data pyramid is determined according to the range of the mine area to be analyzed and the magnetic susceptibility distribution data, the magnetic susceptibility distribution data is decomposed into different levels using the Gaussian pyramid or Laplacian pyramid method, wherein each level corresponds to a different resolution, and the high level retains large-scale features, and the low level retains detailed features;

[0029] At the highest level, the local dynamic threshold is used to preliminarily segment the magnetic susceptibility anomaly and the background area, and layer by layer downward, the boundary of the abnormal area is refined at each level using the local features of the current layer and the segmentation result of the previous layer, and at each level, the gradient operator or Canny operator is used to detect the edge of the abnormal area and the edge information of the upper and lower layers is combined for correction, to obtain the edge detection result of each level;

[0030] An edge map is constructed for each level according to the edge detection result, the edge map is presented in the form of a binary image, and the edge position pixel value is 1 and the remaining positions are 0, the edge maps of each level are weighted and fused to obtain the fused edge map, and the weight coefficients of the edge maps are determined according to the resolution of each level;

[0031] The seed filling algorithm is used to fill the fused edge map into a solid area map, and the holes and noise points in the solid area map are eliminated through the closing operation and the opening operation in the morphological operation, to obtain the identification result of the magnetic susceptibility abnormal area.

[0032] As a preferred aspect of the invention, the specific steps for extracting features from the identified magnetic susceptibility anomaly regions to obtain their structural features, and then inputting these structural features into a trained geological structure classification model to obtain the geological structure classification results for the magnetic susceptibility anomaly regions, are as follows:

[0033] Morphological analysis of the magnetic susceptibility anomaly region yields its geometric characteristics, including its area, volume, aspect ratio, and sphericity. Intensity analysis of the magnetic susceptibility anomaly region yields its intensity characteristics, including the mean, standard deviation, maximum, and minimum values ​​of magnetic susceptibility.

[0034] A geological structure classification model is constructed and trained using machine learning algorithms. The geometric and intensity features of the magnetic susceptibility anomaly region are then input into the trained geological structure classification model to obtain the geological structure classification results of the magnetic susceptibility anomaly region.

[0035] The present invention has the following advantages:

[0036] 1. This invention uses an initial three-dimensional geological model and a layered hybrid inversion method to invert magnetic susceptibility, obtaining magnetic susceptibility distribution data for the mining area to be analyzed. This not only fully leverages the advantages of different inversion algorithms and compensates for the shortcomings of a single algorithm, thereby improving the accuracy of magnetic susceptibility inversion results, but also employs efficient local algorithms and powerful global algorithms for the geological characteristics of shallow and deep areas respectively, improving the computational efficiency and stability of magnetic susceptibility inversion. Furthermore, the set buffer region can smoothly transition and integrate the results of different magnetic susceptibility inversions, ensuring the continuity and reliability of the inversion results and facilitating subsequent magnetic susceptibility anomaly identification. This enhances the inversion accuracy and identification effect of this magnetic susceptibility identification method for deep structural analysis of mining areas.

[0037] 2. This invention identifies anomalous regions of magnetic susceptibility distribution in the mining area by using multi-scale fusion hierarchical segmentation based on magnetic susceptibility distribution data. This results in the identification of anomalous magnetic susceptibility regions. It not only extracts and fuses information from different scales and effectively captures the size and shape of anomalous regions to improve identification accuracy and completeness, but also reduces errors that may arise from single-scale analysis. Furthermore, it highlights important features, enhances noise resistance, and makes the boundaries of anomalous regions clearer. Multi-scale analysis also provides more comprehensive geological information, thus providing a more reliable basis for subsequent resource exploration and development, thereby improving the identification effect of this magnetic susceptibility identification method for deep structural analysis in mining areas.

[0038] 3. This invention extracts features from identified magnetic susceptibility anomaly areas and classifies the structural features of these areas using a geological structure classification model, obtaining geological structure classification results for the magnetic susceptibility anomaly areas. This enables automatic and accurate identification and classification of magnetic susceptibility anomaly areas, significantly improving analysis efficiency and accuracy, thereby reducing human error and providing a scientific basis for deep structural analysis and resource exploration in mining areas. This enhances the identification effect of this magnetic susceptibility identification method for deep structural analysis in mining areas. Attached Figure Description

[0039] Figure 1 This is a flowchart illustrating the magnetic susceptibility identification method used for deep structural analysis in mining areas, as employed in an embodiment of the present invention. Detailed Implementation

[0040] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of this invention.

[0041] Example 1: A method for identifying magnetic susceptibility in deep geological structures of mining areas, such as... Figure 1 As shown, it includes the following steps:

[0042] Data acquisition and preprocessing: Geological data of the mining area to be analyzed is collected, and magnetic susceptibility measurement data is obtained from the surface of the mining area and boreholes through magnetic susceptibility sensors. Some examples of magnetic susceptibility measurement data are shown in Table 1 below. The magnetic susceptibility measurement data is preprocessed to obtain the preprocessed magnetic susceptibility measurement data.

[0043] Model construction and inversion: An initial three-dimensional geological model is constructed based on the geological data of the mining area to be analyzed and the preprocessed magnetic susceptibility measurement data. The magnetic susceptibility is then inverted based on the initial three-dimensional geological model and the layered hybrid inversion method to obtain the magnetic susceptibility distribution data of the mining area to be analyzed.

[0044] Magnetic susceptibility anomaly identification is based on magnetic susceptibility distribution data and hierarchical segmentation through multi-scale fusion to identify anomalous regions of magnetic susceptibility distribution in the mining area to be analyzed, and to obtain the identification results of magnetic susceptibility anomaly regions, including the location, shape and size of the magnetic susceptibility anomaly regions;

[0045] The structure is analyzed and classified. Feature extraction is performed on the identified magnetic susceptibility anomaly regions to obtain their structural features. These features are then input into the trained geological structure classification model to obtain the geological structure classification results for the magnetic susceptibility anomaly regions.

[0046]

[0047] Table 1. Partial examples of magnetic susceptibility measurement data.

[0048] The specific steps for preprocessing the magnetic susceptibility measurement data to obtain preprocessed magnetic susceptibility measurement data are as follows:

[0049] Wavelet transform denoising process is used to decompose the magnetic susceptibility measurement data into high-frequency and low-frequency components. The low-frequency component containing geological information is retained while the high-frequency component with noise is removed. Then, the denoised magnetic susceptibility measurement data is reconstructed by inverse discrete wavelet transform.

[0050] Inverse distance weighted spatial interpolation is used to obtain the position information of each data point in the magnetic susceptibility measurement data, identify the target position to be interpolated, calculate the Euclidean distance between the target position and all known data points, select the power parameter and set the distance weight of all known data points based on the Euclidean distance, and calculate the magnetic susceptibility value of the target position based on the distance weight.

[0051] Moving average smoothing involves selecting a fixed-length sliding window and moving it forward point by point from the beginning of the magnetic susceptibility measurement data sequence. Each time the window moves forward, a subset of data points covered by the current sliding window is selected, and the arithmetic mean of all data points within the current sliding window is calculated. This arithmetic mean is used as the smoothed value corresponding to the center position of the sliding window. The average values ​​calculated for each sliding window are then arranged sequentially to form a new smoothed data sequence.

[0052] The specific steps for constructing the initial three-dimensional geological model based on the geological data of the mining area to be analyzed and the preprocessed magnetic susceptibility measurement data are as follows:

[0053] Collect geological data, including borehole data, geological profiles, and surface geological maps of the mining area to be analyzed.

[0054] The stratigraphic units are divided. Based on the collected geological data of the mining area to be analyzed, the main stratigraphic units of the mining area to be analyzed are divided and the magnetic susceptibility range of each unit is determined.

[0055] Delineate the structural interfaces: Based on the geological profile map in the geological data of the mining area to be analyzed, delineate the main structural interfaces of the mining area to be analyzed, including faults and rock strata contact surfaces.

[0056] Three-dimensional geological modeling: Based on the geological structure information of the mining area obtained above and the preprocessed magnetic susceptibility measurement data, an initial three-dimensional geological model is constructed using three-dimensional modeling software.

[0057] The specific steps for obtaining the magnetic susceptibility distribution data of the mining area to be analyzed by inverting the magnetic susceptibility based on the initial three-dimensional geological model and using the layered hybrid inversion method are as follows:

[0058] The theoretical magnetic susceptibility data for each magnetic susceptibility measurement point is calculated using the finite difference or finite element numerical method and based on the initial three-dimensional geological model. The theoretical magnetic susceptibility data obtained from the forward model calculation is compared with the actual preprocessed magnetic susceptibility measurement data, and the differences between the two are analyzed.

[0059] Based on geological characteristics and depth, the mining area to be analyzed is divided into layers. An objective function is constructed by combining regularization constraints. Local optimization algorithms and global optimization algorithms are selected, and magnetic susceptibility inversion is performed on the layered mining area using a layered hybrid inversion method. The parameters of the initial three-dimensional geological model are adjusted through iterative optimization to minimize the objective function. In each iteration of optimization, forward calculations are performed based on the parameters of the new three-dimensional geological model until the error between the theoretical magnetic susceptibility data and the actual preprocessed magnetic susceptibility measurement data meets the convergence condition, thus obtaining the magnetic susceptibility distribution data of the mining area to be analyzed.

[0060] The specific steps for stratifying the mining area to be analyzed based on geological characteristics and depth, constructing an objective function in conjunction with regularization constraints, selecting local and global optimization algorithms, and performing magnetic susceptibility inversion on the stratified mining area using a stratified hybrid inversion method are as follows:

[0061] Based on geological characteristics and depth, the mining area to be analyzed is divided into shallow region, buffer region and deep region, and an objective function is constructed by combining regularization constraints;

[0062] A local optimization algorithm is selected to perform magnetic susceptibility inversion in the shallow region of the mining area to be analyzed, and a global optimization algorithm is selected to perform magnetic susceptibility inversion in the deep region of the mining area to be analyzed.

[0063] For the buffer zone of the mining area to be analyzed, the magnetic susceptibility is inverted using both local and global optimization algorithms. The results of the two algorithms are obtained. The weights of the magnetic susceptibility inversion results of the two algorithms are set according to the geological depth and then weighted and fused to obtain the magnetic susceptibility inversion result of the buffer zone. The weight of the local optimization algorithm is the ratio of the distance between the data point to be inverted and the deep area to the overall width of the buffer zone, while the weight of the global optimization algorithm is the ratio of the distance between the data point to be inverted and the shallow area to the overall width of the buffer zone.

[0064] It should be noted that the shallow areas of the mining area have relatively simple geological structures, and local optimization algorithms are used for magnetic susceptibility inversion, including algorithms such as the conjugate gradient method. However, for the complex geological structures of the deep areas, global optimization algorithms are used for inversion, including algorithms such as the genetic algorithm. Furthermore, in the buffer zone, the weights of the inversion results are gradually adjusted according to the depth, so that the weights of the shallow algorithm gradually decrease and the weights of the deep algorithm gradually increase. This ensures that the two inversion results achieve good continuity and consistency at the boundary.

[0065] The above steps, based on the initial three-dimensional geological model, invert the magnetic susceptibility using a layered hybrid inversion method to obtain the magnetic susceptibility distribution data of the mining area to be analyzed. This not only fully leverages the advantages of different inversion algorithms and compensates for the shortcomings of a single algorithm, thereby improving the accuracy of the magnetic susceptibility inversion results, but also utilizes efficient local algorithms and powerful global algorithms respectively for the geological characteristics of shallow and deep areas to improve the computational efficiency and stability of magnetic susceptibility inversion. Furthermore, the set buffer region can smoothly transition and integrate the results of different magnetic susceptibility inversions, ensuring the continuity and reliability of the inversion results and facilitating subsequent magnetic susceptibility anomaly identification. This enhances the inversion accuracy and identification effect of this magnetic susceptibility identification method for deep structural analysis of mining areas.

[0066] The specific steps for identifying anomalous regions of magnetic susceptibility based on magnetic susceptibility distribution data and through hierarchical segmentation using multi-scale fusion to obtain the identification results of anomalous magnetic susceptibility regions are as follows:

[0067] Based on the extent of the mining area to be analyzed and the magnetic susceptibility distribution data, the decomposition levels of the data pyramid are determined. The Gaussian pyramid or Laplace pyramid method is used to decompose the magnetic susceptibility distribution data into different levels, each of which corresponds to a different resolution. The higher levels retain large-scale features, while the lower levels retain detailed features.

[0068] At the highest level, local dynamic thresholds are used to initially segment the magnetic susceptibility anomaly and background regions. Layer by layer downwards, at each level, the boundaries of the anomaly region are refined using the local features of the current layer and the segmentation results of the previous layer. At each level, gradient operators or Canny operators are used to detect the edges of the anomaly region and are corrected by combining the edge information of the upper and lower layers to obtain the edge detection results of each level.

[0069] Based on the edge detection results, an edge map is constructed for each level. The edge map is presented in the form of a binary image, where the pixel value at the edge position is 1 and the value at the other position is 0. The edge maps of each level are weighted and fused to obtain the fused edge map. The weight coefficients of the edge map are determined according to the resolution of each level.

[0070] The seed filling algorithm is used to fill the fused edge map into a solid region map, and the holes and noise points in the solid region map are eliminated by the closing and opening operations in morphological operations to obtain the identification results of the magnetic susceptibility abnormal region.

[0071] It should be noted that multi-scale fusion hierarchical segmentation is not only applicable to anomaly recognition in two-dimensional images, but can also be applied to the recognition of magnetic susceptibility anomalies in three-dimensional distributions. Although the processing of three-dimensional data is relatively complex, the basic idea is similar to that of two-dimensional data, that is, to capture anomalous features at different scales through multi-scale decomposition and feature fusion. Furthermore, the solid region map is optimized by using morphological operations to further improve the recognition effect of anomalous regions. This can effectively handle problems such as noise, breaks and burrs in the image, and make the boundaries of anomalous regions smoother and more coherent.

[0072] The above steps, based on magnetic susceptibility distribution data, identify anomalous areas of magnetic susceptibility distribution in the mining area to be analyzed through hierarchical segmentation using multi-scale fusion. This process not only extracts and fuses information from different scales and effectively captures the size and shape of anomalous areas to improve identification accuracy and completeness, but also reduces errors that may be caused by single-scale analysis. Furthermore, it highlights important features, enhances noise resistance, and makes the boundaries of anomalous areas clearer. Moreover, multi-scale analysis provides more comprehensive geological information, thus providing a more reliable basis for subsequent resource exploration and development, thereby improving the identification effect of this magnetic susceptibility identification method for deep structural analysis in mining areas.

[0073] The specific steps for extracting features from the identified magnetic susceptibility anomaly regions to obtain their structural features, and then inputting these structural features into the trained geological structure classification model to obtain the geological structure classification results for the magnetic susceptibility anomaly regions, are as follows:

[0074] Morphological analysis of the magnetic susceptibility anomaly region yields its geometric characteristics, including its area, volume, aspect ratio, and sphericity. Intensity analysis of the magnetic susceptibility anomaly region yields its intensity characteristics, including the mean, standard deviation, maximum, and minimum values ​​of magnetic susceptibility.

[0075] A geological structure classification model is constructed and trained using machine learning algorithms. The geometric and intensity features of the magnetic susceptibility anomaly region are then input into the trained geological structure classification model to obtain the geological structure classification results of the magnetic susceptibility anomaly region.

[0076] The above steps extract features from the identified magnetic susceptibility anomaly areas and classify the structural features of these areas using a geological structure classification model. This yields the geological structure classification results for the magnetic susceptibility anomaly areas, enabling automatic and accurate identification and classification. This significantly improves analysis efficiency and accuracy, reduces human error, and provides a scientific basis for deep structural analysis and resource exploration in mining areas. It also enhances the identification effect of this magnetic susceptibility identification method for deep structural analysis in mining areas.

[0077] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims. Parts not described in detail in this specification are prior art known to those skilled in the art.

Claims

1. A method for identifying magnetic susceptibility in deep geological structures of mining areas, characterized in that, Includes the following steps: Data acquisition and preprocessing: Collect geological data of the mining area to be analyzed and obtain magnetic susceptibility measurement data from the surface of the mining area and boreholes through magnetic susceptibility sensors. Preprocess the magnetic susceptibility measurement data to obtain preprocessed magnetic susceptibility measurement data. Model construction and inversion: An initial three-dimensional geological model is constructed based on the geological data of the mining area to be analyzed and the preprocessed magnetic susceptibility measurement data. Based on the initial three-dimensional geological model, the magnetic susceptibility is inverted using a layered hybrid inversion method to obtain the magnetic susceptibility distribution data of the mining area to be analyzed. The specific steps of the layered hybrid inversion method are as follows: Based on geological characteristics and depth, the mining area to be analyzed is divided into shallow region, buffer region and deep region, and an objective function is constructed by combining regularization constraints; A local optimization algorithm is selected to perform magnetic susceptibility inversion in the shallow region of the mining area to be analyzed, and a global optimization algorithm is selected to perform magnetic susceptibility inversion in the deep region of the mining area to be analyzed. For the buffer zone of the mining area to be analyzed, the magnetic susceptibility of the buffer zone is inverted using local optimization algorithm and global optimization algorithm respectively, and the magnetic susceptibility inversion results of the two algorithms are obtained. The weights of the magnetic susceptibility inversion results of the two algorithms are set according to the geological depth and the two are weighted and fused to obtain the magnetic susceptibility inversion result of the buffer zone. The weight of the local optimization algorithm is the proportion of the distance between the data point to be inverted and the deep area to the overall width of the buffer zone, while the weight of the global optimization algorithm is the proportion of the distance between the data point to be inverted and the shallow area to the overall width of the buffer zone. Magnetic susceptibility anomaly identification is based on magnetic susceptibility distribution data and hierarchical segmentation through multi-scale fusion to identify anomalous regions of magnetic susceptibility distribution in the mining area to be analyzed, and to obtain the identification results of magnetic susceptibility anomaly regions, including the location, shape and size of the magnetic susceptibility anomaly regions; The structure is analyzed and classified. Feature extraction is performed on the identified magnetic susceptibility anomaly regions to obtain their structural features. These features are then input into the trained geological structure classification model to obtain the geological structure classification results for the magnetic susceptibility anomaly regions.

2. The magnetic susceptibility identification method for deep structural analysis in mining areas according to claim 1, characterized in that, The specific steps for preprocessing the magnetic susceptibility measurement data to obtain preprocessed magnetic susceptibility measurement data are as follows: Wavelet transform denoising process is used to decompose the magnetic susceptibility measurement data into high-frequency and low-frequency components. The low-frequency component containing geological information is retained while the high-frequency component with noise is removed. Then, the denoised magnetic susceptibility measurement data is reconstructed by inverse discrete wavelet transform. Inverse distance weighted spatial interpolation is used to obtain the position information of each data point in the magnetic susceptibility measurement data, identify the target position to be interpolated, calculate the Euclidean distance between the target position and all known data points, select the power parameter and set the distance weight of all known data points based on the Euclidean distance, and calculate the magnetic susceptibility value of the target position based on the distance weight. Moving average smoothing involves selecting a fixed-length sliding window and moving it forward point by point from the beginning of the magnetic susceptibility measurement data sequence. Each time the window moves forward, a subset of data points covered by the current sliding window is selected, and the arithmetic mean of all data points within the current sliding window is calculated. This arithmetic mean is used as the smoothed value corresponding to the center position of the sliding window. The average values ​​calculated for each sliding window are then arranged sequentially to form a new smoothed data sequence.

3. The magnetic susceptibility identification method for deep structural analysis in mining areas according to claim 2, characterized in that, The specific steps for constructing the initial three-dimensional geological model based on the geological data of the mining area to be analyzed and the preprocessed magnetic susceptibility measurement data are as follows: Collect geological data, including borehole data, geological profiles, and surface geological maps of the mining area to be analyzed. The stratigraphic units are divided. Based on the collected geological data of the mining area to be analyzed, the main stratigraphic units of the mining area to be analyzed are divided and the magnetic susceptibility range of each unit is determined. Delineate the structural interfaces: Based on the geological profile map in the geological data of the mining area to be analyzed, delineate the main structural interfaces of the mining area to be analyzed, including faults and rock strata contact surfaces. Three-dimensional geological modeling: Based on the geological structure information of the mining area obtained above and the preprocessed magnetic susceptibility measurement data, an initial three-dimensional geological model is constructed using three-dimensional modeling software.

4. The magnetic susceptibility identification method for deep structural analysis in mining areas according to claim 3, characterized in that, The specific steps for obtaining the magnetic susceptibility distribution data of the mining area to be analyzed by inverting the magnetic susceptibility based on the initial three-dimensional geological model and using the layered hybrid inversion method are as follows: The theoretical magnetic susceptibility data for each magnetic susceptibility measurement point is calculated using the finite difference or finite element numerical method and based on the initial three-dimensional geological model. The theoretical magnetic susceptibility data obtained from the forward model calculation is compared with the actual preprocessed magnetic susceptibility measurement data, and the differences between the two are analyzed. Based on geological characteristics and depth, the mining area to be analyzed is divided into layers. An objective function is constructed by combining regularization constraints. Local optimization algorithms and global optimization algorithms are selected, and magnetic susceptibility inversion is performed on the layered mining area using a layered hybrid inversion method. The parameters of the initial three-dimensional geological model are adjusted through iterative optimization to minimize the objective function. In each iteration of optimization, forward calculations are performed based on the parameters of the new three-dimensional geological model until the error between the theoretical magnetic susceptibility data and the actual preprocessed magnetic susceptibility measurement data meets the convergence condition, thus obtaining the magnetic susceptibility distribution data of the mining area to be analyzed.

5. The magnetic susceptibility identification method for deep structural analysis in mining areas according to claim 4, characterized in that, The specific steps for identifying anomalous regions of magnetic susceptibility based on magnetic susceptibility distribution data and through hierarchical segmentation using multi-scale fusion to obtain the identification results of anomalous magnetic susceptibility regions are as follows: Based on the extent of the mining area to be analyzed and the magnetic susceptibility distribution data, the decomposition levels of the data pyramid are determined. The Gaussian pyramid or Laplace pyramid method is used to decompose the magnetic susceptibility distribution data into different levels, each of which corresponds to a different resolution. The higher levels retain large-scale features, while the lower levels retain detailed features. At the highest level, local dynamic thresholds are used to initially segment the magnetic susceptibility anomaly and background regions. Layer by layer downwards, at each level, the boundaries of the anomaly region are refined using the local features of the current layer and the segmentation results of the previous layer. At each level, gradient operators or Canny operators are used to detect the edges of the anomaly region and are corrected by combining the edge information of the upper and lower layers to obtain the edge detection results of each level. Based on the edge detection results, an edge map is constructed for each level. The edge map is presented in the form of a binary image, where the pixel value at the edge position is 1 and the value at the other position is 0. The edge maps of each level are weighted and fused to obtain the fused edge map. The weight coefficients of the edge map are determined according to the resolution of each level. The seed filling algorithm is used to fill the fused edge map into a solid region map, and the holes and noise points in the solid region map are eliminated by the closing and opening operations in morphological operations to obtain the identification results of the magnetic susceptibility abnormal region.

6. The magnetic susceptibility identification method for deep structural analysis in mining areas according to claim 5, characterized in that, The specific steps for extracting features from the identified magnetic susceptibility anomaly regions to obtain their structural features, and then inputting these structural features into the trained geological structure classification model to obtain the geological structure classification results for the magnetic susceptibility anomaly regions, are as follows: Morphological analysis of the magnetic susceptibility anomaly region yields its geometric characteristics, including its area, volume, aspect ratio, and sphericity. Intensity analysis of the magnetic susceptibility anomaly region yields its intensity characteristics, including the mean, standard deviation, maximum, and minimum values ​​of magnetic susceptibility. A geological structure classification model is constructed and trained using machine learning algorithms. The geometric and intensity features of the magnetic susceptibility anomaly region are then input into the trained geological structure classification model to obtain the geological structure classification results of the magnetic susceptibility anomaly region.

Images