A geological exploration data extraction method

Abstract

The present application relates to geological exploration technical field, specifically disclose a kind of geological exploration data extraction method, including multi-source data acquisition, pre-processing, feature data extraction, data fusion and verification and data output steps.Ground penetrating radar, gravity, magnetic data and soil sample are collected;Pre-processing contains filtering, correction and sample processing;Feature extraction contains identifying fault fold, inverting anomaly body distribution, analyzing element content;Fusion data builds correlation model, field verification and accuracy is less than 85% when rework;Finally output the data that verification passes.The present application comprehensively multi-source data, uses adaptive median filtering and other technologies, after fusion verification, the comprehensiveness, accuracy, reliability and efficiency of data extraction are improved, adapt to complex geological environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geological exploration, and particularly refers to a geological exploration data extraction method. BACKGROUND

[0002] Geological exploration is an important means to study the surface and internal geological structure of the earth, the distribution of mineral resources and the like, and the accuracy and comprehensiveness of data extraction directly affect subsequent geological analysis, resource exploration and the like.

[0003] The traditional geological exploration data extraction method often relies on a single data source, such as analyzing only through geological radar data or gravity and magnetic method data, which has the problem of one-sidedness of data and is difficult to comprehensively reflect the underground geological conditions. At the same time, in the data processing process, there is a lack of effective fusion and verification mechanism, resulting in low accuracy of the extracted data, which may mislead the subsequent geological research and resource development.

[0004] In addition, the traditional method has low standardization and automation degree in data processing, and is greatly influenced by human factors, and the data output format is single, which is not convenient for subsequent visual analysis and application.

[0005] Therefore, a geological exploration data extraction method has become a problem to be solved by people. SUMMARY

[0006] The technical problem to be solved by the present application is to provide a geological exploration data extraction method to improve the accuracy and comprehensiveness of geological exploration data extraction.

[0007] To solve the above technical problem, the technical scheme provided by the present application is as follows: a geological exploration data extraction method, comprising the following steps:

[0008] S1, multi-source data acquisition: acquiring radar reflection data of underground geological structure by using a geological radar, collecting gravity data and magnetic method data of the exploration area by a gravimeter and a magnetic method instrument respectively, and collecting soil samples by using a soil sampler;

[0009] S2, data preprocessing: filtering and gain processing the radar reflection data, topographic correction and normal field correction of the gravity data and the magnetic method data, and air-drying, grinding and screening processing of the soil samples;

[0010] S3, feature data extraction: analyzing the radar profile to identify the interruption and dislocation phenomenon of the reflection wave group to determine the position and trend of the fault from the preprocessed radar reflection data, and judging the type and distribution range of the fold according to the bending shape of the reflection wave group;

[0011] An iterative inversion method is used to perform inversion calculation on the gravity data and the magnetic method data, the calculated theoretical gravity value and the magnetic method value are compared with the measured values by constructing an initial geological model, the initial geological model parameters are continuously adjusted until the errors of the two are within the allowable range, and the underground anomaly body distribution characteristics are obtained;

[0012] Element analysis is performed on the soil samples to obtain characteristic element content data;

[0013] S4, data fusion and verification: fuse the extracted various feature data to establish a multi-source data correlation model, select a survey point for field verification and calculate the data extraction accuracy, and if the accuracy is lower than the preset threshold, return to step S2 or S3 for reprocessing;

[0014] The field verification obtains actual underground geological samples through drilling or pit exploration, compares the actual geological sample analysis results with the extracted data to calculate the accuracy; and the preset threshold is 85%.

[0015] S5, data output: the extracted data that passes the verification is stored in a preset format.

[0016] Further, in step S1, the device moving speed is kept uniform during the geological radar scanning process, the scanning interval is set to 5-10 meters; the collection point interval of the gravimeter and the magnetic method instrument is set to 10-50 meters; the sampling depth of the soil sampler is 0-1.5 meters, and at least 500 grams of soil sample is collected at each sampling point.

[0017] Further, in step S2, the radar reflection data is filtered by an adaptive median filtering algorithm to remove noise, and the adaptive median filtering algorithm filters the abnormal values in the data by setting a dynamic window size; the terrain correction of the gravity data and the magnetic method data uses a digital terrain model to calculate and correct the terrain influence according to the elevation of the measurement point and the surrounding terrain characteristics.

[0018] Further, the radar reflection data is filtered by an adaptive median filtering algorithm to remove noise, and the specific method of the adaptive median filtering algorithm for filtering the abnormal values in the data is as follows:

[0019] (1) Set the initial window size to 3x3 and the maximum window size to S max ;

[0020] (2) For each pixel point in the radar reflection data matrix, extract all pixel values in the neighborhood window centered on it with the current window size;

[0021] (3) Calculate the minimum value Z min = min{Z(i,j) | (i,j) ∈ W} and the maximum value Z max= max{Z(i,j) | (i,j) ∈ W} and median Z med = median{Z(i,j) | (i,j) ∈ W}; where W represents the current neighborhood window, Z(i,j) represents the radar reflection data value of pixel point (i,j) in the window;

[0022] (4) If Z min < Z med < Z max , execute step (5), otherwise, execute step (2) again. k+1 = S k + 2 increase the window size and repeat steps (2)-(3) until the window size reaches S max .

[0023] (5) If Z min < Z(x,y) < Z max , output Z(x,y), otherwise, output Z med ; where Z(x,y) is the original radar reflection data value at coordinate (x,y), S max is set to an odd window size not more than 11x11 according to the noise characteristics of radar data.

[0024] Further, in step S3, the specific method for determining the location and strike of the fault and judging the type and distribution range of the fold is as follows:

[0025] Construct a reflection wave group tracking model: perform two-dimensional gridding processing on the preprocessed radar profile data, set the reflection intensity of the i-th survey line and the j-th sampling point in the profile as S(i,j), track the continuous wave group by using a dynamic programming algorithm, and define a wave group continuity index C(i,j):

[0026]

[0027] Wherein σ is the standard deviation of the reflection intensity, when C(i,j) > 0.6max(C), it is determined as the same wave group, and a wave group number matrix G(i,j) is established;

[0028] Fault identification: calculate the wave group position offset Δd(i,j) between adjacent survey lines:

[0029] Δd(i,j) = |arg max(G(i,j)) - arg max(G(i+1,j))|;

[0030] When Δd(i,j) > 3Δx and more than 5 consecutive sampling points satisfy this condition, mark it as a wave group interruption / dislocation area, and the center coordinates (x f , y f ) are the fault location, and Δx is the sampling interval; the strike of the dislocation area is fitted by the least square method, and the strike angle θ satisfies:

[0031]

[0032] where (x k ,y k ) is the point coordinate along the fault line, is the mean value;

[0033] Folding recognition: calculate the curvature parameters for the same wave group:

[0034] where S ′ (i,j) and S″(i,j) are the first and second derivatives of the reflection intensity respectively; when K(i,j)>0 and the continuous wave group presents symmetrical bending, it is determined as an anticline, and when K(i,j)<0 and the continuous wave group presents symmetrical bending, it is determined as a syncline; the folding distribution range is the minimum circumscribed rectangle of the wave group bending area, and the length of the long axis is the folding extension length, and the length of the short axis is the folding width.

[0035] Further, in step S3, the specific method for obtaining the distribution characteristics of the underground anomaly body is as follows:

[0036] 1) Construct an initial geological model: divide the exploration area into N three-dimensional grid units, and the density parameter of each unit is ρ k (k=1,2,...,N), and the magnetization intensity parameter is M k (k=1,2,...,N), to form an initial parameter vector p0=[ρ1,ρ2,...,ρ N ,M1,M2,...,M N ] T ;

[0037] 2) Calculate the theoretical value: calculate the theoretical gravity value and the theoretical magnetic method value based on the initial geological model

[0038] where the theoretical gravity value of the i-th measuring point is:

[0039] where G is the gravitational constant, μ0 is the vacuum permeability, (x i ,y i ,z) is the measuring point coordinate, (x k ,y k ,z k ) is the center coordinate of the k-th grid unit, V k is the volume of the k-th grid unit;

[0040] 3) Calculation of error: define the objective function E(p) as the mean square error between theoretical value and measured value:

[0041]

[0042] where m is the number of measuring points, Δg obs,i , T obs,i are the measured gravity value and magnetic value respectively, σ Δg,i , σ T,i are the standard deviations corresponding to the measured values;

[0043] 4) Iterative optimization: update the model parameters by using the damped least square method, and the parameter vector of the t+1 iteration is:

[0044]

[0045] where J t is the Jacobian matrix at the t iteration, the element or W is the weight matrix, the diagonal element is e t is the residual vector, λ t is the damping factor, and I is the unit matrix;

[0046] 5) Convergence judgment: repeat steps 2) to 4) until the objective function E(p t )≤ε; ε is a preset error threshold, and the value range is 10 -5 -10 -3 ; the model parameters corresponding to the underground anomaly body distribution characteristics at this time are the inversion results.

[0047] Further, in step S3, the element analysis of the soil sample is performed by using the X-ray fluorescence spectrum analysis method to measure the content of the characteristic elements related to the mineral resources in the soil, and the characteristic elements include copper, iron and gold.

[0048] Further, in step S4, the specific method of data fusion and verification is as follows:

[0049] Data standardization processing: normalize the extracted geological structure feature data G, underground anomaly body distribution feature data A and soil characteristic element content data E:

[0050]

[0051]

[0052] where G i , A j , E k are the original feature data, G i′ , A j ′ , E k ′ is the standardized data;

[0053] Constructing a multi-source data correlation model: a weighted fusion algorithm is used to establish a correlation model M, and the output value of the correlation model is:

[0054] M(x, y) = w G ·G ′ (x, y) + w A ·A ′ (x, y) + w E ·E ′ (x, y);

[0055] where w G , w A , w E are weight coefficients, satisfying w G + w A + w E = 1, the weight values are determined by the analytic hierarchy process, and w G ∈ [0.3, 0.4], w A ∈ [0.3, 0.4], w E ∈ [0.2, 0.4];

[0056] Selecting verification points: n exploration verification points are selected by spatial stratified sampling method, and the spatial coordinates of each verification point are (x m , y m ), wherein n ≥ 30 and are uniformly distributed in the exploration area;

[0057] Calculating data extraction accuracy: the actual geological data R m of the verification point is obtained by drilling or pit exploration, the model output value M(x m , y m ) is compared with the actual data R m , and the accuracy P is calculated:

[0058]

[0059] wherein when R m = 0, if M(x m , y m ) = 0, it is counted as accurate, otherwise it is counted as error;

[0060] Threshold judgment and feedback: if P ≥ 85%, it is determined that the data extraction is qualified; if P < 85%, the error contribution rate of each data type is calculated:

[0061]

[0062] When C G >0.4, return to step S2 to reprocess the radar reflection data, when C A >0.4, return to step S3 to reprocess the gravity and magnetic method data, when C E >0.4, return to step S3 to reprocess the soil sample data, otherwise, return to steps S2 and S3 simultaneously for optimization.

[0063] Further, in step S5, the preset format includes an Excel table containing the position and attribute information of each feature data and a Shapefile vector file for visual display and analysis in a GIS system.

[0064] Further, the soil powder obtained by grinding and sieving the soil sample has a particle size less than 0.15 mm.

[0065] Compared with the prior art, the present application has the following advantages:

[0066] The present application comprehensively reflects the geological and mineral information of the exploration area by collecting multi-source data, including geological radar, gravity, magnetic method and soil samples, thereby improving the comprehensiveness of data extraction.

[0067] The present application adopts a series of data preprocessing and feature extraction methods, such as adaptive median filtering, terrain correction, inversion calculation, etc., effectively removing noise and interference in the data, thereby improving the accuracy of data extraction.

[0068] The present application introduces a data fusion and verification step, which further ensures the reliability of the extracted data by correlating multi-source data and field verification, and enhances the adaptability of the method to complex geological environments.

[0069] The present application has a clear process and high degree of automation, reduces manual operation, and improves the efficiency of geological and mineral exploration data extraction. BRIEF DESCRIPTION OF DRAWINGS

[0070] Figure 1 is a flowchart of a geological exploration data extraction method of the present application. DETAILED DESCRIPTION

[0071] The various exemplary embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be noted that the relative arrangement, numerical expressions and values of the components and steps described in these embodiments do not limit the scope of the present application unless otherwise specifically stated.

[0072] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way limiting to the scope of the application and its applications or uses.

[0073] Techniques, methods, and apparatus known to those of ordinary skill in the relevant art can not be discussed in detail, but should be considered part of the specification where appropriate.

[0074] In all of the examples shown and discussed herein, any specific values should be interpreted as merely exemplary, and not as a limitation. Thus, other examples of the exemplary embodiments can have different values.

[0075] A geological exploration data extraction method will be further described in detail below in combination with the accompanying drawings.

[0076] In combination with the accompanying drawings Figure 1 The present application will be described in detail.

[0077] A geological exploration data extraction method, comprising the following steps:

[0078] S1, multi-source data acquisition: radar reflection data of underground geological structure is obtained by using a geological radar, gravity data and magnetic method data of the exploration area are collected by a gravimeter and a magnetic method instrument respectively, and soil samples are collected by using a soil sampler.

[0079] In this step, the collection of multi-source data can obtain underground geological information from different angles. The radar reflection data can reflect the morphological characteristics of the underground geological structure. The gravity and magnetic method data can reflect the physical property differences of the underground material. The soil sample contains geochemical information, which provides a rich data basis for subsequent comprehensive analysis.

[0080] Specifically, the device moving speed is kept uniform during the scanning process of the geological radar, and the scanning interval is set to 5-10 meters, so that the continuity and consistency of the radar reflection data can be ensured, and data distortion caused by changes in moving speed or excessively large or small intervals can be avoided. The point interval of the gravimeter and the magnetic method instrument is set to 10-50 meters. Reasonable point interval can improve data collection efficiency while ensuring data accuracy. The sampling depth of the soil sampler is 0-1.5 meters, and at least 500 grams of soil sample is collected at each sampling point. This depth range can obtain soil information related to surface and shallow geological activity, and sufficient sample amount can meet the subsequent analysis requirements.

[0081] S2, data preprocessing: filtering and gain processing are performed on the radar reflection data; topographic correction and normal field correction are performed on the gravity data and magnetic method data, and the soil samples are air-dried, ground and sieved.

[0082] Data preprocessing is a key link to improve data quality. By processing different types of data, interference factors can be eliminated or weakened, laying a good foundation for subsequent feature extraction.

[0083] For radar reflection data, an adaptive median filtering algorithm is used to remove noise, which filters the outliers in the data by setting a dynamic window size. The specific method is as follows:

[0084] (1) Set the initial window size to 3x3 and the maximum window size to S max ;

[0085] (2) For each pixel point in the radar reflection data matrix, extract all pixel values in the neighborhood window centered on it with the current window size;

[0086] (3) Calculate the minimum value Z min = min{Z(i,j) | (i,j) ∈ W}, the maximum value Z max = max{Z(i,j) | (i,j) ∈ W} and the median value Z med = median{Z(i,j) | (i,j) ∈ W} of the pixel values in the neighborhood window; where W represents the current neighborhood window and Z(i,j) represents the radar reflection data value of pixel point (i,j) in the window;

[0087] (4) If Z min < Z med < Z max , execute step (5), otherwise increase the window size by S k+1 = S k +2 and repeat steps (2)-(3) until the window size reaches S max ;

[0088] (5) If Z min < Z(x,y) < Z max , output Z(x,y), otherwise output Z med ; where Z(x,y) is the original radar reflection data value at coordinate (x,y), S max is set to an odd number not exceeding 11x11 according to the noise characteristics of radar data.

[0089] This adaptive median filtering algorithm can dynamically adjust the window size according to the noise in the data, effectively remove noise while preserving the details of the radar reflection data.

[0090] The terrain correction of gravity data and magnetic data uses a digital terrain model to calculate the terrain influence and correction according to the elevation of the measurement point and the surrounding terrain features, to eliminate the interference of terrain undulations on gravity and magnetic measurement results; normal field correction can remove the influence of the earth's normal gravity field and normal magnetic field, highlighting the anomaly signal.

[0091] The soil samples are air-dried, ground and sieved, and the soil powder obtained after grinding and sieving has a particle size of less than 0.15 mm, so that the uniformity of the soil samples can be ensured and the accuracy of subsequent element analysis can be improved.

[0092] S3, feature data extraction: analyzing the radar profile from the preprocessed radar reflection data to identify the interruption and displacement of the reflection group to determine the position and trend of the fault, and judging the type and distribution range of the fold according to the bending shape of the reflection group; using an iterative inversion method to calculate the gravity data and magnetic data, comparing the calculated theoretical gravity value and magnetic value with the measured value by constructing an initial geological model, and constantly adjusting the initial geological model parameters until the error is within the allowable range to obtain the distribution characteristics of the underground anomaly body; and performing element analysis on the soil samples to obtain characteristic element content data.

[0093] Feature data extraction is the core link of the method, and valuable geological feature information is extracted from the preprocessed data through specific algorithms and methods.

[0094] The specific method for determining the position and trend of the fault and judging the type and distribution range of the fold is as follows:

[0095] Constructing a reflection group tracking model: performing two-dimensional gridding processing on the preprocessed radar profile data, setting the reflection intensity of the i-th measuring line and the j-th sampling point in the profile as S(i,j), tracking the continuous wave group using a dynamic programming algorithm, and defining a wave group continuity index C(i,j):

[0096]

[0097] Where σ is the standard deviation of the reflection intensity, and when C(i,j)>0.6max(C), it is determined that it is the same wave group, and a wave group numbering matrix G(i,j) is established;

[0098] Fault identification: calculating the wave group position offset Δd(i,j) between adjacent measuring lines:

[0099] Δd(i,j)=|arg max(G(i,j))-arg max(G(i+1,j))|;

[0100] When Δd(i,j)>3Δx and more than 5 consecutive sampling points satisfy this condition, it is marked as a wave group interruption / displacement area, and the center coordinates (x f ,y f ) are the fault position, and Δx is the sampling interval; the trend of the displacement area is fitted by the least squares method, and the trend angle θ satisfies:

[0101]

[0102] Where (xk ,y k () represents the coordinates of points along the fault line. The mean;

[0103] Wrinkle identification: Calculate curvature parameters for the same wave group:

[0104] Where S ′ (i,j) and S″(i,j) are the first and second derivatives of the reflection intensity, respectively. When K(i,j)>0 and the continuous wave group is symmetrically curved, it is determined to be an anticline. When K(i,j)<0 and the continuous wave group is symmetrically curved, it is determined to be a syncline. The fold distribution range is the smallest circumscribed rectangle of the wave group's curved region. Its major axis length is the fold extension length, and its minor axis length is the fold width.

[0105] The above methods can accurately identify geological structural features such as faults and folds, providing a reliable basis for geological structural research.

[0106] The specific method for obtaining the distribution characteristics of underground anomalies is as follows:

[0107] 1) Constructing the initial geological model: Divide the exploration area into N three-dimensional grid cells, with the density parameter ρ for each cell. k (k = 1, 2, ..., N), magnetization parameter is M k (k = 1, 2, ..., N), forming an initial parameter vector p0 = [ρ1, ρ2, ..., ρ N ,M1,M2,...,M N ] T ;

[0108] 2) Calculate theoretical values: Calculate theoretical gravity values ​​based on the initial geological model. and theoretical magnetic values The theoretical gravity value of the i-th measuring point is:

[0109] The theoretical magnetic value of the i-th measuring point is:

[0110] Where G is the gravitational constant, μ0 is the free permeability, (x i ,y i (x, z) are the coordinates of the measuring point, (x, z) are the coordinates of the measuring point. k ,y k ,z k () represents the coordinates of the center of the k-th grid cell. V k Let k be the volume of the k-th grid cell;

[0111] 3) Calculation error: Define the objective function E(p) as the mean square error between the theoretical and measured values:

[0112]

[0113] Where m is the number of measurement points, Δg obs,i T obs,i These are the measured gravity value and the magnetic field value, respectively, σ Δg,i σ T,i These are the standard deviations of the corresponding measured values;

[0114] 4) Iterative optimization: The model parameters are updated using the damped least squares method. The parameter vector for the (t+1)th iteration is:

[0115]

[0116] In the formula J t Let be the Jacobian matrix at iteration t, with elements or W is the weight matrix, and its diagonal elements are... e t Let λ be the residual vector. t I is the damping factor, and I is the identity matrix;

[0117] 5) Convergence check: Repeat steps 2)-4) until the objective function E(p) is achieved. t )≤ε; ε is the preset error threshold, with a value range of 10. -5 -10 -3 The distribution characteristics of underground anomalies corresponding to the model parameters at this time are the inversion results.

[0118] Through iterative inversion methods, geological models can be continuously optimized to make theoretical calculations and measured values ​​as close as possible, thereby accurately obtaining the distribution characteristics of underground anomalies and providing important clues for mineral resource exploration.

[0119] X-ray fluorescence spectrometry was used to perform elemental analysis on soil samples. This method has the advantages of fast analysis speed, high accuracy and simple sample pretreatment. It can accurately determine the content of mineral-related characteristic elements in the soil, including copper, iron and gold. Anomalies in the content of these elements can indicate the presence of mineral resources.

[0120] S4. Data Fusion and Verification: The extracted feature data are fused to establish a multi-source data association model. Exploration points are selected for field verification, and the data extraction accuracy is calculated. If the accuracy is lower than a preset threshold, the process returns to step S2 or S3 for reprocessing. The field verification involves obtaining actual underground geological samples through drilling or pit exploration. The analysis results of the actual geological samples are compared with the extracted data to calculate the accuracy. The preset threshold is 85%.

[0121] Data fusion can organically combine characteristic data from different sources, making full use of the advantages of various types of data to improve the integrity and reliability of geological information; data verification can evaluate the accuracy of the extracted results to ensure that they meet the requirements of practical applications.

[0122] The specific methods for data fusion and verification are as follows:

[0123] Data standardization processing: The extracted geological structural feature data G, subsurface anomaly distribution feature data A, and soil characteristic element content data E are normalized.

[0124]

[0125] Among them G i A j E k These are the original feature data, G i ′ A j ′ E k ′ The data is standardized.

[0126] Data standardization can eliminate the dimensional differences between different types of data, making all types of data comparable and facilitating fusion analysis.

[0127] Constructing a multi-source data association model: A weighted fusion algorithm is used to establish an association model M. The output value of the association model is:

[0128] M(x,y)=w G ·G ′ (x,y)+w A ·A ′ (x,y)+w E ·E ′ (x,y);

[0129] Where w G w A w E Let w be the weighting coefficient. G +w A +w E =1, the weight values ​​are determined by the analytic hierarchy process, and w G ∈[0.3,0.4], w A ∈[0.3,0.4], w E ∈[0.2,0.4];

[0130] The analytic hierarchy process (AHP) can reasonably determine the weight coefficients based on the importance of various types of data, enabling the fusion model to more accurately reflect comprehensive geological information.

[0131] Verification point selection: n exploration verification points were selected using spatial stratified sampling. The spatial coordinates of each verification point are (x, y, z). m ,y m ), where n≥30 and are evenly distributed throughout the exploration area;

[0132] The verification points selected by the spatial stratified sampling method are highly representative and can comprehensively reflect the geological conditions of the exploration area, ensuring the reliability of the verification results.

[0133] Calculate the accuracy of data extraction: Obtain the actual geological data R of the verification point through drilling or pitting. m The model output value M(x) m ,y m ) and actual data R m Compare the results and calculate the accuracy P:

[0134]

[0135] Where R m When = 0, if M(x) m ,y m If ) = 0, it is considered accurate; otherwise, it is considered an error.

[0136] Threshold judgment and feedback: If P ≥ 85%, the data extraction is deemed qualified; if P < 85%, the error contribution rate of each data type is calculated.

[0137]

[0138] When C G When C > 0.4, return to step S2 to reprocess the radar reflection data. A When C > 0.4, return to step S3 to reprocess the gravity and magnetism data. E If the value is greater than 0.4, return to step S3 to reprocess the soil sample data; otherwise, return to steps S2 and S3 simultaneously for optimization.

[0139] Error contribution rate analysis can identify the main data types that cause the accuracy to fall short of the target, allowing for targeted reprocessing and optimization to improve the quality of data extraction.

[0140] S5. Data Output: Organize and store the verified extracted data according to the preset format.

[0141] The preset formats include Excel spreadsheets and Shapefile vector files. The Excel spreadsheets contain the location and attribute information of each feature data, which facilitates statistical analysis and management of the data. The Shapefile vector files are used for visualization and analysis in GIS systems, which can intuitively present the spatial distribution of underground geological features and provide convenience for geological research and decision-making.

[0142] The specific implementation process of the geological exploration data extraction method of the present invention is as follows:

[0143] Taking a certain metal mine exploration area (2km × 3km) as the research object, the above-mentioned geological exploration data extraction method was used to carry out data collection and analysis. The specific process is as follows:

[0144] I. Multi-source data acquisition

[0145] Ground-penetrating radar data: The SIR-4000 ground-penetrating radar was used with an antenna frequency of 100MHz. It moved at a constant speed (2m / s) along the survey line with a scanning interval of 8m. A total of 25 survey lines were collected, and the data format was rad format. The number of sampling points was 512 points / survey line.

[0146] Gravity and magnetic data: A CG-5 gravimeter and a GSM-19T magnetometer were used, with a sampling point spacing of 30m to form a 100m×100m grid of measuring points. A total of 600 gravity data (unit: mGal) and 600 magnetic data (unit: nT) were obtained.

[0147] Soil samples: A column sampler was used, with a sampling depth of 0-1.2m. Sampling points were arranged in a 50m×50m grid, and a total of 240 soil samples were collected. Each sample weighed 600g and was placed in a polyethylene sample bag with a number for preservation.

[0148] II. Data Preprocessing

[0149] Radar reflection data processing:

[0150] Adaptive median filtering is used, with an initial window size of 3×3 and S... max Let it be 7×7, σ=15 (derived from statistical data).

[0151] After filtering, the noise signal intensity in the radar profile is reduced by 40%, and the clarity of the reflected wave group is improved.

[0152] The gain processing uses exponential gain to compensate for depth attenuation, with a gain coefficient k = 0.02 / m.

[0153] Gravity and magnetic correction:

[0154] Terrain correction: Based on a 1:10,000 digital terrain model (DEM resolution 5m), terrain correction values ​​were calculated, with the maximum correction values ​​being +3.2mGal (gravity) and -5.6nT (magnetic method).

[0155] Normal field correction: The normal field of gravity adopts the 1980 International Gravity Formula, and the normal field of magnetic field adopts the IGRF2020 model. The range of outliers after correction is: gravity -5.8 to +4.2 mGal, magnetic field -35 to +42 nT.

[0156] Soil sample preparation: After air drying for 72 hours, the soil was ground in an agate mortar and passed through a 100-mesh sieve (0.15 mm aperture) to obtain uniform soil powder, which was then placed in a sample bottle for analysis.

[0157] III. Feature Data Extraction

[0158] Geological structure identification:

[0159] Wave group tracking: Calculate C(i,j), set the threshold to 0.6×max(C)=0.6×28=16.8, and establish an indexing matrix G(i,j) for 5 consecutive wave groups.

[0160] Fault identification: Δx = 0.5m (sampling interval), calculate the position offset Δd(i,j) between adjacent survey lines. A fault is marked when Δd(i,j) > 1.5m and this is satisfied for 5 consecutive sampling points. A total of 3 faults were identified, with fault F1 having center coordinates of (500m, 800m) and a strike angle θ = 35° (fitted using least squares method).

[0161] Wrinkle identification: Calculate the curvature parameter K(i,j). An anticlinal structure (K>0, symmetrical bending) was found in the 12-15 section of the survey line. The distribution range is a rectangular area (x:1200-1800m, y:500-900m), with a major axis of 1200m and a minor axis of 800m.

[0162] Subsurface anomaly inversion:

[0163] Initial model: The region is divided into a 20×30×5 three-dimensional mesh (x: 200m / unit, y: 200m / unit, z: 100m / unit), N = 3000, initial density ρ k =2.67g / cm 3 Magnetization M k =0.01A / m.

[0164] Iterative inversion: Objective function threshold ε = 5 × 10 -4 After 12 iterations and convergence, the distribution of the anomalous body was obtained: a high-density body (ρ = 2.8-3.0 g / cm³) exists in the region (1000-1500 m, 1200-1800 m). 3The presence of high magnetic anomalies (M = 0.05-0.08 A / m) suggests that the mineral is a magnetic ore body.

[0165] Soil elemental analysis:

[0166] The contents of Cu, Fe, and Au were determined using an X-ray fluorescence spectrometer (model XRF-1800). The results showed that in the soil above the anomaly, the Cu content was 120-350 ppm (background value 30-50 ppm), the Fe content was 5.2-8.6% (background value 2.0-3.5%), and the Au content was 0.1-0.5 ppb (background value <0.05 ppb).

[0167] IV. Data Fusion and Validation

[0168] Data standardization:

[0169] Geological structural data G: Fault density (number of faults / km) 2 ), degree of fold development (0-1 index), normalized range 0.2-0.8;

[0170] Anomaly data A: Density anomalies and magnetic anomalies, normalized range 0.3-0.9;

[0171] Elemental data E: Anomaly indices for Cu, Fe, and Au, normalized to range 0.4–0.95;

[0172] Association model construction:

[0173] The Analytic Hierarchy Process (AHP) determines the weights: w G =0.35,w A =0.35, w E =0.3;

[0174] Field verification:

[0175] Thirty verification points were selected using spatial stratified sampling, including 10 drilling verifications (hole depth 50-100m) and 20 pit verifications (depth 5-10m).

[0176] Accuracy calculation: P = 92.3% ≥ 85%, which meets the threshold requirement;

[0177] Each error contribution rate: C G =0.22,C A =0.28, C E =0.50 (not exceeding 0.4), no need to return for reprocessing.

[0178] V. Data Output

[0179] The Excel spreadsheet contains 300 records, with fields including location coordinates (x, y), fault parameters, depth of anomalies, and element content.

[0180] Shapefile files generate vector layers of fault, fold, and anomalous body distributions, with attribute tables associated with feature data, which can be overlaid and displayed in ArcGIS.

[0181] Final results: Using the method of this invention, three main faults and one anticline structure were extracted in the region, and two high-probability mineralization anomaly zones (each with an area of ​​0.8 km²) were delineated. 2 and 1.2km 2 The soil element anomalies are highly consistent with geological structures and geophysical anomalies, providing a reliable basis for the deployment of subsequent exploration projects.

[0182] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method of extracting geological survey data, characterized by: Comprise the following steps: S1, multi-source data acquisition: use ground penetrating radar to obtain radar reflection data of underground geological structure, use gravity meter and magnetic method instrument to collect gravity data and magnetic data of the survey area respectively, and use soil sampler to collect soil samples; S2, data preprocessing: filtering and gain processing of radar reflection data; Topographic correction and normal field correction are performed on gravity data and magnetic data, and soil samples are dried, ground and sieved; S3, feature data extraction: analyze the radar profile to identify the interruption and dislocation of the reflection wave group to determine the position and trend of the fault, and judge the type and distribution range of the fold according to the bending shape of the reflection wave group; The specific method of determining the position and trend of the fault and judging the type and distribution range of the fold is as follows: Constructing the reflection wave group tracking model: the pre-processed radar profile data is two-dimensionally gridded, and the reflection intensity of the first trace in the profile and the first sampling point is , the dynamic programming algorithm is used to track the continuous wave group, and the wave group continuity index is defined: ； wherein is the standard deviation of the reflection intensity, when is determined to be the same wave group, and a wave group number matrix is established ; Fault identification: calculate wave group position offset between adjacent lines : ； When and the condition is met for more than 5 consecutive sampling points, mark the area as a fault interruption / dislocation area, with the center coordinates is the fault position, is the sampling interval; the strike of the dislocation area is fitted by the least square method, and the strike angle is met: ； wherein is the point coordinate along the fault, , is the mean value; FOLD IDENTIFICATION: Calculate curvature parameter for the same wave group: ; wherein , are the first and second derivatives of the reflection intensity, respectively; when and the continuous wave group is symmetrically curved, it is determined to be a backfold, and when and the continuous wave group is symmetrically curved, it is determined to be a synform; the range of the fold distribution is the minimum circumscribed rectangle of the wave group curved area, the length of the long axis is the extension length of the fold, and the length of the short axis is the width of the fold; Iterative inversion method is used to calculate the gravity data and magnetic data, the calculated theoretical gravity value and magnetic value are compared with the measured value, the initial geological model parameters are adjusted until the error is within the allowable range, and the underground anomaly body distribution characteristics are obtained; Element analysis is performed on the soil samples to obtain soil characteristic element content data; S4, data fusion and verification: normalize the extracted geological structure feature data, underground anomaly body distribution characteristic data and soil characteristic element content data, and use weighted fusion algorithm to establish a multi-source data correlation model, select survey points for field verification and calculate the data extraction accuracy, if the accuracy is lower than the preset threshold, return to step S2 or S3 for reprocessing; The field verification obtains the actual underground geological sample by drilling or pit exploration, compares the analysis results of the actual geological sample with the extracted data to calculate the accuracy; the preset threshold is 85%; S5, data output: the verified extraction data is stored in the preset format.

2. The method of claim 1, wherein: In step S1, the device moves uniformly during the scanning process of the ground penetrating radar, the scanning interval is set to 5-10 meters; the sampling point interval of the gravity meter and the magnetic method instrument is set to 10-50 meters; the sampling depth of the soil sampler is 0-1.5 meters, and at least 500 grams of soil sample is collected at each sampling point.

3. The method of claim 2, wherein: In step S2, adaptive median filtering algorithm is used to remove noise from radar reflection data, and the adaptive median filtering algorithm filters the abnormal values in the data by setting dynamic window size; the topographic correction of gravity data and magnetic data uses digital terrain model, and the topographic influence is calculated and corrected according to the elevation of the measurement point and the surrounding terrain characteristics.

4. The method of claim 3, wherein: In step S3, the specific method of obtaining underground anomaly body distribution characteristics is as follows: 1) constructing an initial geological model: dividing the exploration area into three-dimensional grid cells, the density parameter of each cell being , the magnetization parameter being , forming an initial parameter vector ; 2) Calculate the theoretical value: calculate the theoretical gravity value based on the initial geological model and the theoretical magnetic value , wherein the theoretical gravity value of the first measurement point is: ; The theoretical magnetic method value of the first measuring point is: ;​ wherein, is the gravitational constant, is the vacuum permeability, is the measurement point coordinate, is the coordinate of the center of the grid cell, ; is the volume of the grid cell; 3) Calculate error: define objective function as the mean square error of the theoretical and measured values: ； wherein is the number of measurement points, , are the measured gravity values and magnetic values, respectively, , are the standard deviations of the corresponding measured values, respectively. 4) Iterative optimization: update the model parameters using a damped least squares method, the parameter vector of the first iteration is: ​ ； wherein is the Jacobian matrix at the next iteration, the element , is the weight matrix, the diagonal elements of which are , is the residual vector, is the damping factor, is the identity matrix; 5) Convergence judgment: repeat steps 2) - step 4) until the target function ; is a preset error threshold, the value range is ; At this time, the model parameters corresponding to the underground anomaly body distribution characteristics are the inversion results.

5. The method of claim 4, wherein: In step S3, X-ray fluorescence spectroscopy is used for element analysis of soil samples to determine the content of characteristic elements related to minerals in soil, and the characteristic elements include copper, iron and gold.

6. The method of claim 5, wherein: In step S4, the specific method of data fusion and verification is as follows: Data standardization processing: the extracted geological structure feature data , underground anomaly body distribution feature data , soil characteristic element content data is normalized: ； ； ； wherein , , are the original feature data, , , are the standardized data; Constructing multi-source data association model: using weighted fusion algorithm to establish association model , the output value of the association model is: ； Wherein 、 、 Is a weight coefficient, satisfies , the weight value is determined by the analytic hierarchy process, and , , ; Selecting verification points: using spatial stratified sampling method to select verification points, the spatial coordinates of each verification point are , wherein and are uniformly distributed in the exploration area; Compute data extraction accuracy: actual geologic data from a validation point obtained by drilling or pit exploration Compare model output values to actual data and compute accuracy : ； wherein when then it is counted as an error; otherwise it is counted as accurate. then it is counted as an error; otherwise it is counted as accurate. Threshold judgment and feedback: if , it is determined that the data extraction is qualified; if , the error contribution rate of each data type is calculated: ； ； ； When step S2 to reprocess the radar reflection data, when step S3 to reprocess the gravity and magnetic data, when step S3 to reprocess the soil sample data, otherwise return to both steps S2 and S3 for optimization.

7. The method of claim 6, wherein: In step S5, the preset format includes an Excel table and a Shapefile vector file, the Excel table containing position and attribute information of each feature data, and the Shapefile vector file being used for visual display and analysis in a GIS system.

8. The method of claim 7, wherein: The soil sample is ground and sieved to obtain soil powder with a particle size less than 0.15 mm.

Images