Shallow coverage area prospecting method and system based on soil micro-fine particle extraction

Abstract

The invention discloses a prospecting method and a prospecting system for mainly extracting soil micro-fine particles in a shallow coverage area, and the prospecting method comprises the following steps: separating a micro-fine particle part from a soil sample by adopting a wet screening method according to a preset particle size range to obtain a target micro-fine particle sample; the method comprises the following steps: for a target micro-fine particle sample, acquiring chemical component data through an X-ray fluorescence spectrophotometer, and determining the content and distribution characteristics of main elements in the sample; analyzing the mineral composition of the target micro-fine particle sample by adopting an X-ray diffractometer, and identifying the types and relative contents of main minerals in the sample; determining the particle size distribution of the target micro-fine particle sample through a laser particle size analyzer to obtain the proportion information of particles with different particle sizes in the sample; according to the geophysical data, gravity, magnetic force and electrical parameters of the target area are extracted, and a geophysical feature model is established; and carrying out data fusion on the geochemical features and the geophysical feature model, and extracting key feature variables in multi-source information by adopting a principal component analysis method.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method and system for prospecting in shallow coverage areas mainly based on extraction of soil fine particles. Background Art

[0002] In the process of prospecting with soil fine particles, how to obtain high-quality geochemical and mineralogical information is a key technical issue. The extraction and separation of soil fine particles is the first step, but due to the complexity of soil components and the special properties of fine particles, how to ensure that the separated fine particles can truly reflect the geochemical characteristics of the target area is an urgent problem to be solved. In addition, when analyzing the chemical composition, mineral composition and particle size distribution of fine particles, how to select appropriate testing equipment and methods to obtain accurate and reliable data is also a technical challenge. Different testing equipment and methods may affect the results, so it is necessary to optimize the test process and parameter settings for different sample types and test targets. At the same time, in the data interpretation and prospecting prediction stage, how to effectively integrate geological, geophysical and other multi-source information to establish an accurate mineral resource distribution model is also a complex technical problem. It is necessary to consider the characteristics, scales and uncertainties of different data, and select appropriate data fusion methods and modeling algorithms to obtain reliable prediction results. In short, how to cope with technical challenges such as sample complexity, test accuracy and multi-source data integration in the extraction, analysis and testing of soil fine particles and data interpretation is a key problem to be solved in this prospecting method. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for prospecting in shallow coverage areas mainly based on extraction of soil fine particles.

[0004] To achieve the above object, the present invention adopts the following technical solution:

[0005] A prospecting method for shallow overburden areas mainly based on soil fine particle extraction, comprising:

[0006] Separating fine particles from a soil sample to obtain a target fine particle sample;

[0007] For the target fine-grained samples, the chemical composition data is obtained by X-ray fluorescence spectrometer to determine the content of the main elements in the samples and their distribution characteristics;

[0008] Use X-ray diffractometer to analyze the mineral composition of the target fine-grained sample and identify the main mineral types and their relative contents in the sample;

[0009] The particle size distribution of the target fine particle sample is measured by a laser particle size analyzer to obtain the proportion of particles of different particle sizes in the sample;

[0010] Combined with the geological information of the target area, match the chemical composition, mineral composition and particle size distribution data with the geological background to determine whether the geochemical characteristics of the fine-grained samples are consistent with the regional geological characteristics;

[0011] According to the geophysical data, extract the gravity, magnetic and electrical parameters of the target area and establish a geophysical characteristic model;

[0012] Fuse the geochemical characteristics with the geophysical characteristic model, and use the principal component analysis method to extract the key characteristic variables in the multi-source information;

[0013] Based on the key characteristic variables, establish a mineral resource distribution prediction model through the random forest algorithm to obtain the mineral resource distribution probability map of the target area.

[0014] Preferably, the step of combining the geological information of the target area, matching the chemical composition, mineral composition and particle size distribution data with the geological background, and determining whether the geochemical characteristics of the fine-grained samples are consistent with the regional geological characteristics includes:

[0015] Use the geological information of the target area to obtain geological background data;

[0016] According to the geological background data, extract the chemical composition and mineral composition data;

[0017] Generate the distribution characteristic values of the fine-grained samples through the particle size distribution data;

[0018] Integrate the chemical composition, mineral composition and particle size distribution data, and calculate the geochemical characteristic values;

[0019] Match the geochemical characteristic values with the geological background data to generate a matching score;

[0020] If the matching score is higher than the preset threshold, it is determined that the geochemical characteristics are consistent with the regional geological characteristics;

[0021] Based on the matching results, generate an analysis report on the consistency between the geochemical characteristics and geological characteristics of the target area.

[0022] Preferably, the step of extracting the gravity, magnetic and electrical parameters of the target area according to the geophysical data and establishing a geophysical characteristic model includes:

[0023] Obtain the geophysical measurement data of the target area, extract the gravity measurement values, magnetic measurement values and electrical measurement values, and calculate the gravity anomaly values, magnetic anomaly values and electrical parameter values;

[0024] For the gravity anomaly values, magnetic anomaly values and electrical parameter values, respectively construct a gravity characteristic distribution model, a magnetic characteristic distribution model and an electrical characteristic distribution model;

[0025] Integrate the gravity feature distribution model, magnetic feature distribution model, and electrical property feature distribution model to generate a comprehensive geophysical feature model;

[0026] According to the comprehensive geophysical feature model, extract the distribution values of geological bodies, and judge the types and spatial positions of geological bodies;

[0027] If the type of geological body is consistent with the preset type, generate the distribution characteristic values of the geological body;

[0028] According to the distribution characteristic values of the geological body, establish a relationship model between the geological body and geophysical features;

[0029] Based on the relationship model, generate the analysis result of the consistency between geophysical features and geological features in the target area.

[0030] Preferably, for the data fusion of geochemical features and geophysical feature models, the principal component analysis method is used to extract the key feature variables in multi-source information, including:

[0031] Obtain geochemical measurement data and geophysical feature models, input the data into the principal component analysis method for standardization processing to eliminate the dimension difference;

[0032] Use the principal component analysis method to reduce the dimension of the standardized data, calculate the data covariance matrix, and obtain the eigenvalues and eigenvectors;

[0033] Sort according to the eigenvalue size, select the first several principal components to generate a set of principal component feature vectors;

[0034] For the selected principal component feature vectors, reconstruct the geochemical and geophysical feature data to obtain a reduced-dimensional feature set;

[0035] According to the reduced-dimensional feature set, construct a geochemical and geophysical feature fusion model and calculate the feature weights;

[0036] If the feature weights meet the preset threshold, determine the effectiveness of the feature fusion model, otherwise adjust the number of principal components and recalculate;

[0037] Based on the effective geochemical and geophysical feature fusion model, extract the key feature variables and output the feature variable set.

[0038] Preferably, based on the key feature variables, a mineral resource distribution prediction model is established through the random forest algorithm to obtain the mineral resource distribution probability map of the target area, including:

[0039] Use the eigenvalue to reconstruct the variable set, construct a training set based on the data volume, divide the spatial domain for the target area, and implement modeling using the random forest algorithm to obtain the prediction result of mineral resource distribution;

[0040] If the weight value meets the preset conditions, determine the reliability of the predictive model and output the mineral resource distribution probability map; otherwise, adjust the number of eigenvalue and retrain.

[0041] Divide grid cells in the spatial domain, calculate the mineral resource distribution probability of each cell based on the modeling results, and generate the mineral resource distribution probability map of the target area.

[0042] Use the clustering algorithm to divide the probability map into regions and determine the high-probability mineral resource distribution regions.

[0043] Extract the boundary of the concentrated area of the mineral resource distribution according to the spatial data of the high-probability region and generate the mineral resource distribution hotspot map.

[0044] For the mineral resource distribution hotspot map, use the spatial interpolation algorithm to supplement the missing data and optimize the continuity of the mineral resource distribution prediction results.

[0045] Based on the optimized mineral resource distribution prediction results, output the mineral resource distribution probability map and hotspot map of the target area to form the final mineral resource distribution prediction results.

[0046] Preferably, according to the preset particle size range, use the wet sieving method to separate the fine particle part from the soil sample to obtain the target fine particle sample.

[0047] The present invention also provides a prospecting system mainly for extracting soil fine particles in shallow overburden areas, including:

[0048] A fine particle separation module for separating the fine particle part from the soil sample by the wet sieving method according to the preset particle size range to obtain the target fine particle sample.

[0049] A chemical composition analysis module for obtaining chemical composition data of the target fine particle sample through an X-ray fluorescence spectrometer and determining the main element content and its distribution characteristics in the sample.

[0050] A mineral composition analysis module for analyzing the mineral composition of the target fine particle sample by an X-ray diffractometer and identifying the main mineral types and their relative contents in the sample.

[0051] A particle size distribution measurement module for measuring the particle size distribution of the target fine particle sample by a laser particle size analyzer to obtain the proportion information of particles with different particle sizes in the sample.

[0052] A geological information matching module for combining the geological information of the target area and matching the chemical composition, mineral composition, and particle size distribution data with the geological background to determine whether the geochemical characteristics of the fine particle sample are consistent with the regional geological characteristics.

[0053] A geophysical feature extraction module, which is used to extract gravity, magnetic, and electrical parameters of a target area based on geophysical data and establish a geophysical feature model;

[0054] A data fusion module, which is used to perform data fusion on geochemical features and the geophysical feature model, and extract key feature variables in multi-source information by using the principal component analysis method;

[0055] A mineral resource prediction module, which is used to establish a mineral resource distribution prediction model based on key feature variables through the random forest algorithm and obtain a mineral resource distribution probability map of the target area.

[0056] In the present invention, fine particles in soil samples are separated by the wet sieving method, and their chemical composition, mineral composition, and particle size distribution data are obtained by means of X-ray fluorescence spectroscopy, X-ray diffraction, and laser particle size analysis. Combining with regional geological information, the consistency between the geochemical characteristics of fine particle samples and the geological background is judged. At the same time, geophysical parameters of the target area are extracted to establish a feature model, and data fusion is performed with geochemical characteristics. The principal component analysis method is used to extract key feature variables, and finally a mineral resource distribution prediction model is constructed by the random forest algorithm to generate a mineral resource distribution probability map of the target area. The present invention effectively integrates multi-source geological information, improves the accuracy and reliability of mineral resource prediction, and provides a scientific basis for mineral exploration. Description of the Drawings

[0057] Figure 1 It is a flowchart of the prospecting method mainly based on the extraction of soil fine particles in the shallow overburden area of the present invention. Detailed Embodiments

[0058] Next, the technical solutions of the present invention will be described clearly and completely in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0059] Embodiment 1:

[0060] As Figure 1 shown, the embodiment of the present invention provides a prospecting method mainly based on the extraction of soil fine particles in the shallow overburden area, including:

[0061] Step S101, according to a preset particle size range, separate the fine particle part from the soil sample by using the wet sieving method to obtain a target fine particle sample.

[0062] Obtain soil samples and screen the samples using the wet screening method. Determine the particle size partitioning conditions during the screening process based on the preset particle size range. Separate the fine particles in the soil sample by screening. Determine whether the separated fine particles meet the target particle size range. If they meet the target particle size range, obtain the target fine particle sample. If they do not meet the target particle size range, adjust the screening conditions and separate again. Determine the particle size distribution data of the target fine particle sample based on the final separation result.

[0063] Specifically, wet screening of soil samples is a commonly used particle size analysis technique. Through water-assisted screening, soil particles of different sizes can be effectively separated. When performing wet screening, it is first necessary to determine the target particle size range according to the research purpose, for example, to separate clay particles with a particle size of less than 0.02 mm. Then select a sieve with an appropriate aperture, such as 0.02 mm, 0.05 mm, etc., and stack them in order from large to small. Put the soil sample into the top sieve, rinse with water and stir gently to fully disperse the particles. The water flow will carry fine particles through the sieve, while larger particles will remain on the sieve. This process needs to be repeated until the outflowing water becomes clear. The separated particles of each level are dried and weighed at 60°C to obtain the mass fraction of each particle size range. If it is found that the separated fine particles do not meet expectations, such as the 0.02 mm sieve still contains a lot of coarse particles, the screening conditions need to be adjusted. You can try to extend the screening time, increase the water flow intensity, or use ultrasonic assisted dispersion. Sometimes it is necessary to add a dispersant, such as sodium hexametaphosphate solution, to break the bond between particles. Through repeated adjustment and sieving, samples with the target particle size range can be obtained. Based on this, the physical and chemical properties of the soil can be further analyzed. For example, soils with a high content of fine particles usually have strong water retention but poor permeability; while soils with a high content of coarse particles are the opposite. These characteristics have an important impact on agricultural production and engineering construction. In practical applications, some challenges may be encountered. For example, some clay minerals are very easy to disperse in water, which may cause sieve blockage. At this time, alternative methods such as vibrating screens or wet laser particle size analyzers can be considered. In addition, soils with high organic matter content may form agglomerates during wet sieving, affecting the separation effect. In this case, the sample can be treated with hydrogen peroxide first, and then sieved after removing the organic matter. The particle size distribution data obtained by wet sieving can be plotted as a cumulative curve or frequency distribution diagram. These graphs intuitively show the composition characteristics of soil particles and help soil classification and property evaluation. For example, the steepness of the particle size distribution curve reflects the uniformity of the soil, while the shape of the curve implies the characteristics of the sedimentary environment. In general, wet sieving is a simple and effective method for soil particle size analysis. By carefully controlling the sieving conditions and combining other auxiliary techniques, high-quality fine particle samples and reliable particle size distribution data can be obtained. This information is of great value for research and practice in fields such as soil science and environmental engineering.

[0064] In step S102, for the target fine-grained sample, obtain chemical composition data through an X-ray fluorescence spectrometer to determine the main element content and its distribution characteristics in the sample.

[0065] Use an X-ray fluorescence spectrometer to scan the target fine-grained sample to obtain the fluorescence spectrum data of the sample. According to the fluorescence spectrum data, extract the characteristic peak information of the main elements in the sample. Through the characteristic peak information, determine the types of the main elements in the sample and their corresponding wavelength ranges. According to the wavelength ranges, calculate the relative content values of the main elements in the sample. Through the content values, analyze the distribution characteristics of the main elements in the sample. If the element distribution characteristics meet the preset conditions, generate a chemical composition distribution map of the sample. According to the chemical composition distribution map, determine the chemical composition data of the target fine-grained sample.

[0066] Specifically, X-ray fluorescence spectrometry is a commonly used elemental analysis technique that analyzes the elemental composition of a sample by exciting inner-shell electrons of atoms to generate characteristic X-rays. In the analysis of soil fine particles, this method can quickly and non-destructively obtain the elemental information of the sample. Taking a soil fine particle sample from a farmland as an example, first, the sample is placed in an X-ray fluorescence spectrometer for scanning. The instrument emits X-rays to irradiate the sample, exciting the atoms in the sample to generate characteristic fluorescence. This fluorescence is received by the detector and converted into an electrical signal, finally forming fluorescence spectral data. The obtained spectral data usually contains multiple peaks, and each peak corresponds to an element. For example, the Kα characteristic peak of sulfur may appear at 2.31 keV, and the Kα peak of calcium may be at 3.69 keV. By analyzing the positions and intensities of these characteristic peaks, the main elements present in the sample and their relative contents can be determined. Suppose the following main elements are detected in this soil sample: silicon, aluminum, iron, calcium, potassium, and magnesium. For each element, its position in the spectrum can be determined according to the energy range of its characteristic peak. For example, the Kα peak of silicon is usually around 1.74 keV, while the Kα peak of iron is around 6.40 keV. By comparing the intensities of the characteristic peaks of each element, their relative contents can be calculated. Suppose the analysis results show that: silicon accounts for 45%, aluminum accounts for 25%, iron accounts for 15%, calcium accounts for 8%, potassium accounts for 5%, and magnesium accounts for 2%. These data reflect the approximate distribution of each element in the sample. When analyzing the elemental distribution characteristics, some interesting patterns may be found. For example, a relatively high silicon-aluminum ratio may imply that the soil is rich in clay minerals; a relatively high iron content may be related to the red color of the soil; the calcium and magnesium contents may affect the alkalinity of the soil. If these characteristics meet the preset research objectives, a chemical composition distribution map can be further generated. The chemical composition distribution map can visually display the spatial distribution of each element in the sample. For example, it may be found that some elements are enriched at the edges of the fine particles, while others are evenly distributed. This distribution pattern may reflect certain geological or environmental factors during the soil formation process. Through this detailed elemental analysis, not only can the chemical composition data of soil fine particles be obtained, but also a deeper understanding of the properties and formation history of the soil can be achieved. This is of great significance for fields such as agricultural production, environmental monitoring, and geological research. For example, in agriculture, understanding the elemental composition of soil fine particles can help farmers better select suitable crops and fertilizers; in environmental monitoring, the pollution status can be evaluated by tracking the content changes of certain specific elements; in geological research, the elemental composition of fine particles may provide important clues about rock weathering and soil evolution.

[0067] Step S103: Analyze the mineral composition of the target fine particle sample using an X-ray diffractometer to identify the main mineral types and their relative contents in the sample.

[0068] The target fine-grained sample is scanned using an X-ray diffractometer to obtain the diffraction pattern data of the sample. Based on the diffraction pattern data, the characteristic peak information of the minerals in the sample is extracted. Through the characteristic peak information, the types of the main minerals in the sample and their corresponding diffraction angles are determined. Based on the diffraction angles, the relative content values of the main minerals in the sample are calculated. If the relative content values meet the preset conditions, a mineral composition distribution map of the sample is generated. According to the mineral composition distribution map, the spatial distribution characteristics of the minerals in the sample are analyzed. Through the mineral distribution characteristics, the symbiotic relationship of the minerals in the sample is judged. A mineral composition analysis report of the target fine-grained sample is generated.

[0069] Specifically, an X-ray diffractometer is an important instrument for studying the crystal structure of materials. It obtains the structural information of samples by measuring the diffraction phenomenon of X-rays in crystals. When scanning a target fine-grained sample, X-rays are incident on the sample surface at different angles, diffract with crystal planes, and form characteristic diffraction patterns. Taking a rock sample containing quartz, feldspar, and mica as an example, the diffraction pattern obtained by scanning will show the characteristic diffraction peaks unique to these minerals. Quartz has a strong characteristic peak at about 26.6° in the 2θ angle, feldspar has multiple characteristic peaks in the range of 27 - 28°, and mica has an obvious basal plane diffraction peak near 8 - 10°. By analyzing the positions and intensities of these characteristic peaks, the main mineral species present in the sample can be determined. The calculation of the relative content of minerals usually uses the full peak intensity method or the reference intensity ratio method. Taking the full peak intensity method as an example, assuming the main diffraction peak intensities of quartz, feldspar, and mica are 1000, 800, and 600 (relative units) respectively, then their relative contents are approximately 41.7%, 33.3%, and 25%. This calculation method is simple and fast, but the difference in absorption coefficients of minerals needs to be considered. The generated mineral composition distribution map can visually display the spatial distribution of each mineral in the sample. For example, it may be observed that quartz grains are evenly distributed throughout the sample, while feldspar and mica show zonal or massive distributions. Such distribution characteristics may reflect temperature and pressure changes or later alteration during the rock formation process. By analyzing the spatial distribution characteristics of minerals, the symbiotic relationship between minerals can be inferred. For example, if feldspar and mica are often closely connected, while quartz is relatively independently distributed, this may imply that feldspar and mica were formed under similar geological conditions, while quartz may have crystallized or deposited at different stages. The mineral composition analysis report should not only contain quantitative data but also be interpreted in combination with the geological background of the sample. For example, a high content of quartz may indicate a sedimentary rock with a high degree of maturity, while the abundant presence of feldspar and mica may indicate a magmatic origin or weak weathering. This comprehensive analysis can provide important bases for further geological research, such as inferring tectonic environments and analyzing sediment source areas. Through this systematic X-ray diffraction analysis, the composition and structural characteristics of rocks can be revealed from a microscopic perspective, providing key information for understanding geological processes and mineral resource exploration. The advantage of this method is that it can quickly and accurately identify the mineral components in complex mixtures, especially for those fine-grained samples that are difficult to identify by traditional optical methods.

[0070] Step S104, measure the particle size distribution of the target fine-grained sample by a laser particle size analyzer to obtain the proportion information of particles with different particle sizes in the sample.

[0071] The target sample is scanned using a laser particle size analyzer to obtain the scattered light intensity values of the sample. Based on the scattered light intensity values, the scattering characteristic points of particles with different particle sizes in the sample are extracted. Through the scattering characteristic points, the particle size values and distribution values of particles with different particle sizes in the sample are calculated. Based on the particle size values and distribution values, a particle size distribution curve graph of the sample is generated. According to the particle size distribution curve graph, the distribution state values of particles with different particle sizes in the sample are analyzed. Through the distribution state values, the aggregation state characteristics of the particles in the sample are judged. According to the aggregation state characteristics, a particle size distribution analysis report form of the target sample is generated.

[0072] Specifically, a laser particle size analyzer is a precision particle size measurement instrument, and its working principle is based on the laser diffraction principle. When a laser beam irradiates particles dispersed in a liquid, scattered light is generated. By measuring the intensity distribution of the scattered light, the size and distribution of the particles can be calculated. In practical applications, first, the sample is dispersed in an appropriate liquid, such as water or ethanol. Then, the sample solution is injected into the sample cell of the instrument. The laser beam passes through the sample cell and interacts with the particles to generate scattered light. The detector array of the instrument captures this scattered light and converts it into an electrical signal. Taking a mineral sample as an example, suppose there is a powder sample from a certain mining area. After placing the sample in the laser particle size analyzer for measurement, a series of scattered light intensity values are obtained. These data reflect the scattering characteristics of particles with different particle sizes in the sample. Through a special algorithm, the particle size information of the particles can be extracted from these scattering characteristics. For example, it may be found that there are three main particle size distribution peaks in the sample: one is around 5μm, one is around 20μm, and another is around 50μm. These information can help understand the size distribution of particles in the sample. Next, a particle size distribution curve graph can be generated based on these data. This graph usually uses the particle size as the abscissa and the particle volume percentage as the ordinate. By observing the shape of the curve, the distribution state of the particles in the sample can be intuitively understood. For example, if the curve shows a single-peak distribution, it indicates that the particle sizes in the sample are relatively uniform; if it shows a multi-peak distribution, it means that there are multiple different sizes of particles in the sample. After analyzing the particle size distribution curve graph, the distribution state values of particles with different particle sizes in the sample can be obtained. These values may include the median particle size (D50), average particle size, maximum particle size, etc. For example, it may be found that the D50 value of this sample is 15μm, which means that 50% of the particles are smaller than 15μm. Through these distribution state values, the aggregation state characteristics of the particles in the sample can be further judged. If it is found that a large number of particles are aggregated within a certain specific particle size range, it may mean that there is an agglomeration phenomenon in the sample. This situation is relatively common in some mineral samples, especially when the sample contains clay minerals. Finally, based on all these analysis results, a detailed particle size distribution analysis report form can be generated. This report not only contains specific data, such as the percentage distribution in each particle size range, but may also include explanations and speculations on the sample characteristics. For example, if there are a large number of fine particles in the sample, it may mean that the mineral has experienced a long weathering process. These information is of great guiding significance for subsequent mineral processing and utilization. Through this detailed particle size analysis, the physical properties of the sample can be better understood, providing an important basis for subsequent research and applications. Whether in the fields of geological exploration, materials science or environmental science, particle size analysis is an essential basic work.

[0073] Step S105: Combine the geological information of the target area, match the chemical composition, mineral composition, and particle size distribution data with the geological background, and determine whether the geochemical characteristics of the fine-grained samples are consistent with the regional geological characteristics.

[0074] Use the geological information of the target area to obtain geological background data. According to the geological background data, extract the chemical composition and mineral composition data. Through the particle size distribution data, generate the distribution characteristic values of the fine-grained samples. Integrate the chemical composition, mineral composition, and particle size distribution data, and calculate the geochemical characteristic values. Match the geochemical characteristic values with the geological background data to generate a matching score. If the matching score is higher than the preset threshold, it is determined that the geochemical characteristics are consistent with the regional geological characteristics. Based on the matching results, generate an analysis report on the consistency between the geochemical characteristics and geological characteristics of the target area.

[0075] Specifically, geological background data is the basis for studying the geochemical characteristics of the target area. Obtaining geological background data usually includes information such as rock types, stratigraphic ages, and tectonic environments. For example, during a geological survey in a certain mining area, it was found that the area is mainly composed of granite, metamorphic rocks, and sedimentary rocks, among which granite accounts for about 60%, metamorphic rocks account for 30%, and sedimentary rocks account for 10%. These rocks were formed in different geological periods, reflecting the complex geological evolution history of the area. Extracting chemical composition and mineral composition data from geological background data is a key step for further analysis. Taking granite as an example, its main chemical components include silicon dioxide, aluminum oxide, potassium oxide, etc. The mineral composition includes quartz, feldspar, mica, etc. The content of these components can be accurately determined by X-ray fluorescence spectrometer (XRF) and X-ray diffractometer (XRD). The characteristic values of the particle size distribution of fine-grained samples reflect the degree of rock weathering and sedimentary environment. For example, after laser particle size analysis of a river sediment sample, the median particle size obtained is 0.05 mm, the sorting coefficient is 1.2, and the skewness is 0.3. These data indicate that the sample is medium-sorted fine sand and may be derived from a river sedimentary environment with short-distance transportation. Integrating chemical composition, mineral composition, and particle size distribution data, geochemical characteristic values can be calculated. These characteristic values include element content, element ratio, mineral index, etc. For example, the lanthanum / samarium ratio of a sample is calculated to be 5.2, and the iron / titanium ratio is 8.7. These ratios can indicate the genetic type and evolution degree of the rock. Matching the calculated geochemical characteristic values with the geological background data can generate a matching score. The matching score reflects the degree of consistency between the sample characteristics and the regional geological characteristics. For example, if the rare earth element distribution pattern of the sample is highly similar to the typical pattern of regional granite, the matching score may reach 90 points (out of 100). This high matching indicates that the sample is likely to be derived from granite within the region. When the matching score is higher than a preset threshold (such as 85 points), it can be judged that the geochemical characteristics are consistent with the regional geological characteristics. This consistency means that the geochemical characteristics of the sample can represent the regional geological background, providing a reliable basis for further geological exploration and resource assessment. The consistency analysis report generated based on the matching results can comprehensively reflect the relationship between the sample and the regional geological characteristics. The report may include element content anomaly maps, mineral combination comparison tables, geochemical zoning maps, etc. These charts visually display the similarities and differences between the sample characteristics and the regional background, helping to identify potential mineralized areas or environmental anomalies. By systematically applying this analysis process, the efficiency of geological exploration can be improved, the exploration risk can be reduced, and a scientific basis can be provided for mineral resource development and environmental protection.

[0076] Step S106, according to the geophysical data, extract the gravity, magnetic, and electrical parameters of the target area, and establish a geophysical characteristic model.

[0077] Obtain geophysical measurement data of the target area, extract gravity measurement values, magnetic measurement values, and electrical property measurement values, and calculate gravity anomaly values, magnetic anomaly values, and electrical property parameter values. For the gravity anomaly values, magnetic anomaly values, and electrical property parameter values, construct a gravity characteristic distribution model, a magnetic characteristic distribution model, and an electrical property characteristic distribution model respectively. Integrate the gravity characteristic distribution model, the magnetic characteristic distribution model, and the electrical property characteristic distribution model to generate a comprehensive geophysical characteristic model. According to the comprehensive geophysical characteristic model, extract the geological body distribution values, and judge the geological body type and spatial position. If the geological body type is consistent with the preset type, generate geological body distribution characteristic values. Based on the geological body distribution characteristic values, establish a relationship model between the geological body and the geophysical characteristics. Based on the relationship model, generate the analysis result of the consistency between the geophysical characteristics and the geological characteristics of the target area.

[0078] Specifically, obtaining geophysical measurement data of the target area is a crucial step in geological exploration. These data include gravity, magnetic, and electrical measurement values, which reflect the characteristics of the underground geological structure. By processing these raw data, gravity anomaly values, magnetic anomaly values, and electrical parameter values can be calculated. For example, in the exploration of a certain mining area, gravity measurement shows a local high gravity anomaly, which may imply the existence of a high-density ore body underground. Based on these anomaly values, constructing a distribution model is an important means to understand the underground structure. The gravity characteristic distribution model may present a spherical or ellipsoidal high-density area, indicating the possible location of the ore body. The magnetic characteristic distribution model may show a strong magnetic anomaly, suggesting the presence of iron ore. The electrical characteristic distribution model may reveal a low resistivity area, indicating the existence of an aquifer or a conductive ore body. Integrating these models to form a comprehensive geophysical characteristic model can provide a more comprehensive image of the underground structure. For example, in the exploration of a certain oil and gas field, the comprehensive model may show an area with low gravity, low magnetism but high resistivity, which may indicate the existence of a porous rock structure containing oil and gas. Extracting the geological body distribution values from the comprehensive model helps to judge the type and spatial location of the geological body. For instance, in a certain metal mining area, if an anomaly body with high density, high magnetism and low resistivity is found, it is very likely to be an ore body containing metal sulfide. By comparing with the preset types, it can be confirmed whether it conforms to the expected mineralization type. After generating the geological body distribution characteristic values, it is crucial to establish a relationship model between the geological body and the geophysical characteristics. This relationship model can reveal how the physical properties of the geological body affect the geophysical field. For example, in a certain iron ore mining area, it may be found that the volume of the iron ore body is positively correlated with the intensity of the magnetic anomaly and also has a certain relationship with the magnitude of the gravity anomaly. Finally, based on these relationship models, the analysis results of the consistency between the geophysical characteristics and the geological characteristics of the target area can be generated. This result can guide subsequent exploration work and improve exploration efficiency. For example, in a certain copper ore exploration project, if the geophysical characteristics are highly consistent with the characteristics of the known copper deposit, then drilling work can be preferentially arranged to verify this discovery. This method can not only reduce exploration risks, but also optimize resource allocation and improve exploration success rates.

[0079] Step S107: Perform data fusion on the geochemical characteristics and the geophysical characteristic model, and use the principal component analysis method to extract the key characteristic variables in the multi-source information.

[0080] Obtain geochemical survey data and geophysical characteristic models, input the data into the principal component analysis method for standardization to eliminate the dimension difference. Use the principal component analysis method to reduce the dimension of the standardized data, calculate the data covariance matrix, and obtain the eigenvalues and eigenvectors. Sort according to the eigenvalue size, select the first several principal components, and generate the principal component eigenvector set. For the selected principal component eigenvectors, reconstruct the geochemical and geophysical characteristic data to obtain the reduced-dimension characteristic set. According to the reduced-dimension characteristic set, construct a geochemical and geophysical characteristic fusion model and calculate the characteristic weights. If the characteristic weights meet the preset threshold, determine the effectiveness of the characteristic fusion model; otherwise, adjust the number of principal components and recalculate. Based on the effective geochemical and geophysical characteristic fusion model, extract the key characteristic variables and output the characteristic variable set.

[0081] Specifically, the integration of geochemical and geophysical characteristics is a crucial step in the prospecting process, which can improve the accuracy of mineral exploration by combining different types of data. First, obtain geochemical measurement data and geophysical characteristic models. For example, in a gold ore exploration project, the geochemical data includes the contents of elements such as gold, silver, and copper in the soil, while the geophysical characteristic model contains information such as gravity anomalies and magnetic anomalies. To eliminate the dimensional differences between different data, the principal component analysis method is used for standardization. Suppose the gold element content ranges from 0 to 100 ppm, and the gravity anomaly value ranges from -50 to 50 mGal. After standardization, both are transformed into a standard normal distribution with a mean of 0 and a variance of 1. This can ensure that different characteristics have equal importance in subsequent analyses. Next, perform dimensionality reduction on the standardized data. By calculating the covariance matrix, eigenvalues and eigenvectors are obtained. For example, it may be found that the first three principal components explain 85% of the total variance of the data. Select these three principal components as the set of eigenvectors, which not only retains most of the information but also significantly reduces the data dimension. Reconstruct the data using the selected principal component eigenvectors to obtain the dimensionality-reduced feature set. In the gold ore exploration case, it may be found that the first principal component is highly correlated with the gold and silver contents, the second principal component reflects the gravity anomaly characteristics, and the third principal component may represent the magnetic anomaly information. This dimensionality reduction not only reduces the data volume but also reveals potential geological associations. When constructing a feature fusion model, calculate the weights of each feature. Suppose it is found that the weight of the gold content is 0.4, the weight of the gravity anomaly is 0.3, the weight of the magnetic anomaly is 0.2, and the weights of other features are smaller. If these weights meet the preset threshold (such as the weight of the main feature should be greater than 0.2), the model is considered effective. Otherwise, the number of principal components needs to be adjusted or the data quality needs to be re-evaluated. Finally, extract key feature variables based on the effective fusion model. In gold ore exploration, it may be concluded that the gold content, gravity anomaly, and magnetic anomaly are the three most critical variables. The feature set composed of these variables provides a solid foundation for subsequent mineral prediction. The advantage of this fusion method is that it can comprehensively utilize multi-source data and reduce the one-sidedness that may be brought by a single data type. Through dimensionality reduction, not only the complex data structure is simplified, but also the most explanatory features are highlighted, which helps geologists more accurately judge the potential ore-forming areas. At the same time, the weight analysis reveals the relative importance of different geochemical and geophysical characteristics in mineral prediction, providing a scientific basis for the formulation of exploration strategies.

[0082] Step S108, based on the key feature variables, establish a mineral resource distribution prediction model through the random forest algorithm to obtain a mineral resource distribution probability map of the target area.

[0083] The eigenvalue reconstruction variable set is adopted, the training set is constructed based on the data volume, the spatial domain is divided for the target area, and the random forest algorithm is used to implement the modeling to obtain the prediction result of the distribution of mineral resources. If the weight value meets the preset conditions, the reliability of the predictive model is determined, and the probability map of the distribution of mineral resources is output. Otherwise, the number of eigenvalues is adjusted and retrained. Grid cells are divided in the spatial domain, and the distribution probability of mineral resources in each cell is calculated based on the modeling results to generate the probability map of the distribution of mineral resources in the target area. The clustering algorithm is used to divide the probability map into regions to determine the regions with high-probability distribution of mineral resources. Based on the spatial data of the high-probability regions, the boundaries of the concentrated regions of the distribution of mineral resources are extracted to generate the hot spot map of the distribution of mineral resources. For the hot spot map of the distribution of mineral resources, the spatial interpolation algorithm is used to supplement the missing data to optimize the continuity of the prediction result of the distribution of mineral resources. Based on the optimized prediction result of the distribution of mineral resources, the probability map and the hot spot map of the distribution of mineral resources in the target area are output to form the final prediction result of the distribution of mineral resources.

[0084] Specifically, eigenvalue reconstruction of variable sets is a dimensionality reduction technique that can retain the most important information in the data. In mineral resource exploration, there may be hundreds of geochemical and geophysical variables, which can be reduced to dozens of the most representative features through reconstruction. For example, in a copper mine exploration project, the original data contained more than 200 variables, and after reconstruction, 30 main features were retained, which greatly simplified the subsequent modeling process. The random forest algorithm is a commonly used machine learning method that is particularly suitable for processing high-dimensional data and nonlinear relationships. In mineral prediction, it can effectively capture the complex relationship between geological features and mineralization. For example, in a gold mine prediction project, the random forest model comprehensively considered multiple factors such as topography, lithology, and structure, and successfully identified several high-potential areas, two of which were verified in subsequent drilling. Spatial domain partitioning is the process of dividing the study area into several sub-areas. Doing so can better capture local features and improve prediction accuracy. In a large iron ore exploration project, researchers divided an area of ​​5,000 square kilometers into 100 10×10 km grids, modeled each grid separately, and finally obtained a more accurate prediction result. Clustering algorithms are used to identify areas with similar characteristics, which can help divide potential mineralization zones in mineral prediction. For example, in a tungsten mine survey, the K-means clustering algorithm was used to divide the predicted probability map into three categories: high, medium, and low. The high-probability area is highly consistent with the distribution of known mineral deposits, providing important guidance for further exploration. Spatial interpolation is an important method to fill in data gaps. In mineral exploration, due to the limitation of sampling points, it is often necessary to use interpolation technology to infer the situation of unsampled areas. Kriging is a commonly used geostatistical interpolation method that takes into account spatial autocorrelation and can give an uncertainty estimate of the interpolation results. In a lead-zinc ore prediction project, ordinary kriging was used to interpolate the prediction results and generate a continuous mineralization potential distribution map, which provides a reliable basis for exploration decisions. The mineral resource distribution hotspot map intuitively shows the spatial distribution of high-potential areas. In practical applications, this map is often superimposed and analyzed with other geological information. For example, in a copper-gold polymetallic mining area, the predicted hotspot map was combined with information such as structural linearity and magmatic rock distribution to identify several structural-magmatic composite favorable areas, one of which was found to have a large deposit in subsequent exploration. Through this series of steps, key information can be extracted from massive amounts of geological data, a prediction model can be established, and intuitive prediction results can be generated. This method not only improves exploration efficiency and reduces exploration risks, but also provides a scientific basis for the sustainable development of mineral resources.

[0085] Embodiment 2:

[0086] The present invention also provides a prospecting system for shallow overburden areas mainly based on soil fine particle extraction, comprising:

[0087] A fine particle separation module, which is used to separate the fine particle part from the soil sample by wet sieving method according to the preset particle size range to obtain the target fine particle sample;

[0088] A chemical composition analysis module, which is used to obtain chemical composition data for the target fine particle sample through an X-ray fluorescence spectrometer to determine the main element content and its distribution characteristics in the sample;

[0089] A mineral composition analysis module, which is used to analyze the mineral composition of the target fine particle sample by an X-ray diffractometer to identify the main mineral types and their relative contents in the sample;

[0090] A particle size distribution measurement module, which is used to measure the particle size distribution of the target fine particle sample by a laser particle size analyzer to obtain the proportion information of particles with different particle sizes in the sample;

[0091] A geological information matching module, which is used to combine the geological information of the target area to match the chemical composition, mineral composition and particle size distribution data with the geological background to determine whether the geochemical characteristics of the fine particle sample are consistent with the regional geological characteristics;

[0092] A geophysical feature extraction module, which is used to extract the gravity, magnetic and electrical parameters of the target area according to geophysical data to establish a geophysical feature model;

[0093] A data fusion module, which is used to fuse the geochemical characteristics with the geophysical feature model and extract the key feature variables in the multi-source information by principal component analysis method;

[0094] A mineral resource prediction module, which is used to establish a mineral resource distribution prediction model by random forest algorithm based on the key feature variables to obtain the mineral resource distribution probability map of the target area.

[0095] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present application without departing from the spirit and scope of the embodiments of the present application. Thus, if these modifications and variations of the embodiments of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these changes and modifications.

Claims

1. A method for prospecting in shallow cover areas mainly based on extraction of soil fine particles, characterized in that: include: Separating fine particles from a soil sample to obtain a target fine particle sample; For the target fine-grained samples, the chemical composition data is obtained by X-ray fluorescence spectrometer to determine the content of the main elements in the samples and their distribution characteristics; Use X-ray diffractometer to analyze the mineral composition of the target fine-grained sample and identify the main mineral types and their relative contents in the sample; The particle size distribution of the target fine particle sample is measured by a laser particle size analyzer to obtain the proportion of particles of different particle sizes in the sample; Combined with the geological information of the target area, the chemical composition, mineral composition and particle size distribution data are matched with the geological background to determine whether the geochemical characteristics of the fine-grained samples are consistent with the regional geological characteristics; Based on geophysical data, the gravity, magnetic and electrical parameters of the target area are extracted to establish a geophysical characteristic model; The geochemical characteristics and geophysical characteristics models are integrated, and the key characteristic variables in multi-source information are extracted using principal component analysis. Based on key characteristic variables, a mineral resource distribution prediction model is established through the random forest algorithm to obtain the mineral resource distribution probability map of the target area.

2. The method for prospecting in shallow covered areas based on extraction of soil fine particles as claimed in claim 1, characterized in that: The chemical composition, mineral composition and particle size distribution data are matched with the geological background in combination with the geological information of the target area to determine whether the geochemical characteristics of the fine-grained samples are consistent with the regional geological characteristics, including: Use geological information of the target area to obtain geological background data; Extract chemical composition and mineral composition data based on geological background data; Generate distribution characteristic values ​​of fine particle samples through particle size distribution data; Integrate chemical composition, mineral composition and grain size distribution data to calculate geochemical characteristic values; Matching geochemical characteristic values ​​with geological background data to generate a matching score; If the matching score is higher than the preset threshold, the geochemical characteristics are judged to be consistent with the regional geological characteristics; Based on the matching results, a consistency analysis report of the geochemical and geological characteristics of the target area is generated.

3. The method for prospecting in shallow covered areas based on extraction of soil fine particles as claimed in claim 2, characterized in that: The method of extracting the gravity, magnetic and electrical parameters of the target area based on the geophysical data and establishing a geophysical characteristic model includes: Obtain geophysical measurement data of the target area, extract gravity measurement values, magnetic measurement values ​​and electrical measurement values, and calculate gravity anomaly values, magnetic anomaly values ​​and electrical parameter values; According to the gravity anomaly values, magnetic anomaly values ​​and electrical parameter values, the gravity characteristic distribution model, magnetic characteristic distribution model and electrical characteristic distribution model are constructed respectively; Integrate the gravity feature distribution model, magnetic feature distribution model and electrical feature distribution model to generate a comprehensive model of geophysical features; According to the comprehensive model of geophysical characteristics, the distribution value of geological bodies is extracted to determine the type and spatial location of geological bodies; If the geological body type is consistent with the preset type, the geological body distribution characteristic value is generated; According to the distribution characteristic values ​​of geological bodies, the relationship model between geological bodies and geophysical characteristics is established; Based on the relationship model, the consistency analysis results of the geophysical characteristics and geological characteristics of the target area are generated.

4. The method for prospecting in shallow covered areas based on extraction of soil fine particles as claimed in claim 3, characterized in that: The geochemical characteristics and geophysical characteristics models are fused and the key characteristic variables in the multi-source information are extracted by principal component analysis, including: Obtain geochemical measurement data and geophysical feature models, and input the data into principal component analysis for standardization to eliminate dimensional differences; The principal component analysis method is used to reduce the dimension of the standardized data, calculate the data covariance matrix, and obtain the eigenvalues ​​and eigenvectors; Sort by eigenvalues, select the first several principal components, and generate a set of principal component eigenvectors; For the selected principal component eigenvectors, geochemical and geophysical characteristic data are reconstructed to obtain a feature set after dimensionality reduction; Based on the feature set after dimensionality reduction, a geochemical and geophysical feature fusion model is constructed and feature weights are calculated; If the feature weight meets the preset threshold, the validity of the feature fusion model is determined, otherwise the number of principal components is adjusted and recalculated; Based on an effective geochemical and geophysical feature fusion model, key characteristic variables are extracted and a set of characteristic variables is output.

5. The method for prospecting in shallow covered areas based on extraction of soil fine particles as claimed in claim 4, characterized in that: Based on the key characteristic variables, a mineral resource distribution prediction model is established through a random forest algorithm to obtain a mineral resource distribution probability map of the target area, including: The variable set is reconstructed using eigenvalues, a training set is constructed based on the amount of data, the spatial domain is divided for the target area, and the random forest algorithm is used to implement modeling to obtain the distribution prediction results of mineral resources; If the weight value meets the preset conditions, the reliability of the predictive model is determined and the mineral resource distribution probability map is output; otherwise, the number of eigenvalues ​​is adjusted and retraining is performed; Divide grid cells in the spatial domain, calculate the mineral resource distribution probability of each cell based on the modeling results, and generate a mineral resource distribution probability map of the target area; Clustering algorithms are used to divide the probability map into regions to determine the distribution areas of high-probability mineral resources; Based on the spatial data of high-probability areas, the boundaries of concentrated areas of mineral resource distribution are extracted to generate a hotspot map of mineral resource distribution; For the mineral resource distribution hotspot map, a spatial interpolation algorithm is used to supplement missing data and optimize the continuity of the mineral resource distribution prediction results; Based on the optimized mineral resource distribution prediction results, the mineral resource distribution probability map and heat map of the target area are output to form the final mineral resource distribution prediction results.

6. The method for prospecting in shallow covered areas based on extraction of soil fine particles as claimed in claim 5, characterized in that: According to the preset particle size range, the wet sieving method is used to separate the fine particles from the soil sample to obtain the target fine particle sample.

7. A prospecting system based on extraction of soil fine particles in shallow coverage areas, characterized in that: include: A fine particle separation module is used to separate fine particles from a soil sample by wet sieving according to a preset particle size range to obtain a target fine particle sample; The chemical composition analysis module is used to obtain chemical composition data of the target fine particle sample through an X-ray fluorescence spectrometer to determine the content of the main elements in the sample and its distribution characteristics; The mineral composition analysis module is used to analyze the mineral composition of the target fine-grained sample using an X-ray diffractometer and identify the main mineral types and their relative contents in the sample; The particle size distribution measurement module is used to measure the particle size distribution of the target fine particle sample through a laser particle size analyzer to obtain the proportion information of particles of different particle sizes in the sample; The geological information matching module is used to match the chemical composition, mineral composition and particle size distribution data with the geological background in combination with the geological information of the target area, and to determine whether the geochemical characteristics of the fine-grained samples are consistent with the regional geological characteristics; The geophysical feature extraction module is used to extract the gravity, magnetic and electrical parameters of the target area based on the geophysical data and establish a geophysical feature model; Data fusion module, used to fuse geochemical characteristics with geophysical characteristic models, and extract key characteristic variables from multi-source information using principal component analysis; The mineral resource prediction module is used to establish a mineral resource distribution prediction model based on key characteristic variables through the random forest algorithm to obtain a mineral resource distribution probability map of the target area.

Images