Ore prospecting method and system based on rock fragmentation degree characterization tectonic stress

Abstract

The invention provides a prospecting method and system based on rock fragmentation degree characterization tectonic stress, relates to the technical field of geological exploration, and aims to solve the problem of low prediction precision of a mining area in the prior art. Comprising the following steps: acquiring image data and sample data of a to-be-predicted region; constructing a first inversion model based on the sample data, and performing inversion on the to-be-predicted region based on the first inversion model and the hyperspectral image to obtain a spectral index of a target geologic feature factor of the to-be-predicted region; the target geological characteristic factors comprise crystallinity, alteration component concentration and weathering degree; performing data fusion on the crack features extracted from the high spatial resolution image and the spectral index of the target geologic feature factor to obtain a feature matrix; inputting the feature matrix into a second inversion model, and determining a target mining area of the to-be-predicted area; the second inversion model is a tectonic stress field inversion model pre-constructed based on the graph convolutional neural network. Therefore, the prediction precision of the mining area can be improved.

Description

Technical Field

[0001] The present invention relates to the field of geological prospecting technology, and in particular to a prospecting method and system for characterizing tectonic stress based on the degree of rock fragmentation. Background Art

[0002] In mineral resource exploration, the study of tectonic stress field is an important part of revealing the laws of mineralization. Traditional mining area inversion methods often rely on a single data source or qualitative analysis, making it difficult to achieve accurate inversion of complex tectonic stress fields and quantitative prediction of ore-enriched areas.

[0003] In areas with complex tectonic backgrounds and alteration zone characteristics, there is an urgent need for a high-precision mineralization prediction method that integrates multi-source data. Summary of the invention

[0004] The object of the present invention is to provide a prospecting method and system for characterizing tectonic stress based on the degree of rock fragmentation, so as to improve the prediction accuracy of mineralization areas.

[0005] In order to achieve the above object, the present invention provides the following technical solutions: In a first aspect, the present invention provides a prospecting method for characterizing tectonic stress based on rock fragmentation degree, comprising: acquiring image data and sample data of a region to be predicted, wherein the image data comprises a high spatial resolution image and a hyperspectral image, and the sample data is data corresponding to rock samples in a sample area in the region to be predicted; A first inversion model is constructed based on the sample data, and the area to be predicted is inverted based on the first inversion model and the hyperspectral image to obtain a spectral index of a target geological characteristic factor of the area to be predicted; the target geological characteristic factor includes crystallinity, alteration component concentration and weathering degree; Performing data fusion on the fracture characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor to obtain a characteristic matrix; The characteristic matrix is ​​input into a second inversion model to determine the target mineralization area of ​​the area to be predicted; the second inversion model is a tectonic stress field inversion model pre-constructed based on a graph convolutional neural network.

[0006] Optionally, the sample data includes X-ray data and spectral reflectance data obtained based on a short-wave infrared spectrometer; Constructing a first inversion model based on the sample data includes: Based on the spectral reflectance data, construct a spectral library; Calculating the index of the target geological characteristic factor based on the X-ray data; Extracting a target band whose exponential correlation with the target geological characteristic factor is greater than or equal to a preset value from the spectral library, and constructing a spectral index of the target geological characteristic factor based on the target band; The first inversion model is determined based on the spectral index of the target geological characteristic factor.

[0007] Optionally, the X-ray data at least includes X-ray diffraction data obtained based on an X-ray diffraction instrument; The step of calculating the index of the target geological characteristic factor based on the X-ray data comprises: Determining an X-ray diffraction pattern based on the X-ray diffraction data; Based on the X-ray diffraction pattern, the formula is used: ; The crystallinity index is calculated; wherein, is the index of the crystallinity; is the area under a specific diffraction peak; is the total diffraction signal area; is the background signal; 1 is the starting angle; 2 is the end angle; is the incident angle; is the diffraction angle; is the lower limit of the integral; The upper limit of points.

[0008] Optionally, the X-ray data further includes X-ray fluorescence data obtained based on an X-ray fluorescence instrument; The step of calculating the index of the target geological characteristic factor based on the X-ray data comprises: Based on the X-ray fluorescence data, an index of the alteration component concentration and an index of the weathering degree are calculated.

[0009] Optionally, determining the index of the concentration of the alteration component based on the X-ray fluorescence data comprises: Analyze and obtain target elements of alteration components in the rock sample; the alteration components include at least one or more of chlorite, muscovite and pyrite; the target elements include one or more of Fe, Mg, Al and K; Extracting the concentration of the target element from the X-ray fluorescence data; Using the formula: ; The index of the concentration of the alteration component is calculated; wherein, is an index of the concentration of the alteration component, is the concentration of the target element; is the background concentration of the target element.

[0010] Optionally, determining the weathering degree index based on the X-ray fluorescence data comprises: Extracting the content of target oxides from the X-ray fluorescence data; the target oxides include easily weathered oxides and stable oxides; The ratio of the content of the easily weathered oxide to the content of the stable oxide is determined as the index of the weathering degree.

[0011] Optionally, inputting the characteristic matrix into a second inversion model to determine a target mineralization area of ​​the area to be predicted comprises: Inputting the characteristic matrix into the second inversion model to obtain the tectonic stress field distribution of the area to be predicted; Based on the distribution of the tectonic stress field and the preset mapping benchmark, a mineralization potential mapping unit is generated using a GIS tool; the mineralization potential mapping unit includes a high potential area, a medium potential area and a low potential area; The high potential area is determined as the target mineralization area.

[0012] Optionally, the process of extracting the crack characteristics includes: Processing the high spatial resolution image using a preset deformable convolution and a preset directional filter to obtain a crack feature map; The crack characteristics are extracted from the crack characteristic map; the crack characteristics include crack density, crack direction and crack length.

[0013] Compared with the prior art, the present invention provides a prospecting method for characterizing tectonic stress based on the degree of rock fragmentation, obtains image data and sample data of the area to be predicted, constructs an inversion model using the sample data, and inverts the area to be predicted to obtain target indexes corresponding to the crystallinity, concentration of altered components and degree of weathering, thereby solving the problem of low data reliability caused by using only the single data source of image data in the prior art. Then, the crack characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor are fused to obtain a characteristic matrix. Finally, the characteristic matrix is ​​input into a tectonic stress field inversion model based on a graph convolutional neural network to determine the target mineralization area of ​​the area to be predicted. In this way, the output result of the tectonic stress field inversion model is more accurate, thereby improving the prediction accuracy of the mineralization area in the area to be predicted.

[0014] In a second aspect, the present invention further provides a prospecting system for characterizing tectonic stress based on rock fragmentation degree, comprising: a data acquisition module and a data processor; the data acquisition module is in communication connection with the data processor; The data acquisition module includes a drone; the drone is equipped with a high-definition camera and a hyperspectral camera; the high-definition camera is used to collect high-spatial-resolution images of the area to be predicted, and the hyperspectral camera is used to collect hyperspectral images of the area to be predicted; The data processor is used to obtain image data and sample data of the area to be predicted, wherein the image data includes a hyperspectral image and a high spatial resolution image, and the sample data is data corresponding to rock samples in a sample area in the area to be predicted; A first inversion model is constructed based on the sample data, and the area to be predicted is inverted based on the first inversion model and the hyperspectral image to obtain a spectral index of a target geological characteristic factor of the area to be predicted; the target geological characteristic factor includes crystallinity, alteration component concentration and weathering degree; Performing data fusion on the fracture characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor to obtain a characteristic matrix; The characteristic matrix is ​​input into a second inversion model to determine the target mineralization area of ​​the area to be predicted; the second inversion model is a tectonic stress field inversion model pre-constructed based on a graph convolutional neural network.

[0015] Optionally, the data acquisition module further includes a short-wave infrared spectrometer, an X-ray diffraction instrument and an X-ray fluorescence instrument; The short-wave infrared spectrometer is used to analyze and obtain spectral reflectance data of rock samples in the sample area; The X-ray diffraction instrument is used to collect X-ray diffraction data of the rock sample; The X-ray fluorescence instrument is used to collect X-ray fluorescence data of the rock sample. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 One of the flow diagrams of a prospecting method for characterizing tectonic stress based on the degree of rock fragmentation provided by an embodiment of the present invention; Figure 2 A schematic diagram of states of rock samples with different degrees of fragmentation provided by one embodiment of the present invention; Figure 3A second flow chart of a prospecting method for characterizing tectonic stress based on rock fragmentation degree provided by an embodiment of the present invention; Figure 4 An indication map of a target mineralization area provided for one embodiment of the present invention; Figure 5 One of the structural schematic diagrams of a prospecting system for characterizing tectonic stress based on the degree of rock fragmentation provided by an embodiment of the present invention; Figure 6 The second structural schematic diagram of a mineral prospecting system for characterizing tectonic stress based on the degree of rock fragmentation provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to clearly describe the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, words such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and effects. For example, the first threshold and the second threshold are only used to distinguish different thresholds, and their order is not limited. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit them to be different.

[0018] It should be noted that, in the present invention, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the present invention should not be interpreted as being more preferred or advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way. In the present invention, "at least one" means one or more, and "more than one" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist.

[0019] like Figure 1 As shown, an embodiment of the present invention provides a prospecting method for characterizing tectonic stress based on the degree of rock fragmentation, which may include: Step 110: Obtain image data and sample data of the area to be predicted, where the image data includes high spatial resolution images and hyperspectral images, and the sample data is data corresponding to rock samples in the sample area in the area to be predicted.

[0020] It is understandable that before obtaining sample data and image data, the following steps need to be performed: (1) According to the degree of rock fragmentation in the rock sample, provide clear classification standards for the degree of fragmentation: for example, high degree of fragmentation (close to the fault zone layer), medium degree of fragmentation and low degree of fragmentation (for example, intact surrounding rock).

[0021] According to the actual survey, several small areas are selected as sample areas in the area to be predicted to collect rock samples. Rock samples include but are not limited to: ① Broken rock samples: including severely broken to slightly broken rocks. ② Surrounding rock samples: select complete surrounding rocks that are not significantly affected by crushing. ③ Altered mineral samples: collected from the center of the altered zone or the broken zone, closely related to hydrothermal activity such as chlorite, muscovite, etc. ④ Weathered mineral samples: collected from the surface of the rock mass or the weathering crust, directly exposed to the atmospheric environment such as limonite, etc. ⑤ Non-mineral samples. Rock samples need to cover at least the altered zone and the broken zone. It is understandable that the collection of non-mineral samples can ensure that the samples are representative.

[0022] Combination Figure 2 Sorting out the classification and characteristics of rock samples: High fragmentation: crack density is high and the directions are diverse.

[0023] Moderate fragmentation: The cracks are sparse and the direction is somewhat consistent.

[0024] Low chipping: Few or almost no cracks.

[0025] After collecting rock samples, the samples are numbered, classified, and geological background information is recorded to establish mapping units. The classification standards for mapping units are as follows: High potential mapping unit: Stress field characteristics: high stress concentration, manifested as densely distributed fractures, diverse crack directions, and the formation of obvious stress concentration areas.

[0026] Degree of fragmentation: The rock is significantly fragmented, the density of cracks in the fragmentation unit is large and the width is obvious, the filling is enriched with mineralization characteristics, and the altered minerals (such as chlorite and muscovite) are widely distributed.

[0027] Mineralization anomaly: significantly high mineral concentration, high enrichment of alteration components, low chemical weathering index (high degree of weathering), showing strong mineralization indications.

[0028] Medium potential mapping unit: Stress field characteristics: moderate stress distribution, moderate crack density, and relatively consistent crack direction, showing a moderate stress field response.

[0029] Degree of fragmentation: The rock is moderately fragmented. The number of cracks in the fragmentation unit is limited but visible. There are local fillings and mineralization, and the altered minerals are sparsely distributed.

[0030] Mineralization anomaly: The mineral concentration and alteration component enrichment are moderate, the chemical weathering index is in a moderate range (showing certain weathering characteristics), and the mineralization potential is indicative to a certain extent.

[0031] Low potential mapping units: Stress field characteristics: The stress field distribution is weak, the cracks are sparse and have a single direction, and the stress concentration effect within the mapping unit is not significant.

[0032] Degree of fragmentation: The rock is basically intact, with only a few cracks in the fragmentation unit, no obvious filling features, and few or absent altered minerals.

[0033] Mineralization anomaly: low mineral concentration, poor alteration components, high chemical weathering index (low degree of weathering), and no obvious signs of mineralization.

[0034] (2) Use drones to collect image data. First, the drone is equipped with a high-definition camera to collect high-spatial resolution images of the area to be predicted. Second, the drone is equipped with a hyperspectral camera to collect hyperspectral images of the area to be predicted.

[0035] Step 120: construct a first inversion model based on the sample data, and invert the area to be predicted based on the first inversion model and the hyperspectral image to obtain the spectral index of the target geological characteristic factors of the area to be predicted; the target geological characteristic factors include crystallinity, alteration component concentration and weathering degree.

[0036] Specifically, crystallinity affects spectral characteristics: for materials with high crystallinity, their internal atomic arrangement is regular, and lattice vibration is relatively regular, and they usually show sharp and obvious characteristic absorption peaks or emission peaks in the spectrum. For materials with low crystallinity, however, there are more defects and disorder in the atomic arrangement, and the spectral peaks are often relatively broadened and weakened, and some characteristic peaks may even become unclear.

[0037] After obtaining the first inversion model, the relevant information of the crystallinity of the material can be obtained through quantitative analysis of the hyperspectral image. For example, in the X-ray diffraction spectrum, the diffraction peak of the sample with high crystallinity is sharp and high in intensity, and the half-height width is narrow. According to the intensity, width and other parameters of the diffraction peak, the ratio of the area of ​​the diffraction peak to the integral area of ​​the entire diffraction spectrum is used to quantify the crystallinity of the mineral.

[0038] Alternatively, the Scherrer formula can be used to estimate the crystal grain size, and then infer the degree of crystallinity. It is understandable that in infrared spectroscopy and Raman spectroscopy, the crystallinity of a substance can also be qualitatively or semi-quantitatively analyzed by the shape and intensity ratio of characteristic peaks.

[0039] The concentration of alteration components changes the spectral characteristics: Different alteration components have specific spectral absorption or emission characteristics. When the concentration of alteration components changes, the corresponding spectral signal intensity will also change. For example, iron oxides in rocks are common alteration components. As the concentration of iron oxides increases, the absorption valley of the rock spectrum will deepen and the reflectivity will decrease in the visible-near infrared band, especially at wavelengths related to iron ion electronic transitions, such as 0.45μm, 0.52μm, 0.63μm, etc.

[0040] Establishing a quantitative relationship model between spectral characteristic parameters and alteration component concentrations, that is, the first inversion model mentioned above (such as a multivariate linear regression method), can invert the concentration of alteration components. In this embodiment, by acquiring hyperspectral data, the distribution range and relative concentration of altered minerals can be effectively identified and delineated.

[0041] The degree of weathering causes spectral changes: As the degree of weathering increases, the composition and structure of rocks or minerals will change, causing changes in spectral characteristics. In the early stages of weathering, the mineral particles on the surface of the rock begin to break and decompose, which may be manifested in the spectrum as a weakening of the intensity or position shift of some characteristic peaks. In the later stages of weathering, the rock may form a large number of secondary minerals, such as clay minerals, and new absorption features related to secondary minerals will appear in the spectrum. For example, after granite weathers, clay minerals such as kaolin are formed, and absorption peaks related to hydroxyl groups in kaolin will appear in the near-infrared spectrum.

[0042] Spectral analysis of weathering degree: The weathering degree can be analyzed and judged based on the spectral characteristics. By extracting characteristic parameters in the spectrum, such as spectral slope, absorption peak depth, reflectivity ratio, etc., and comparing and statistically analyzing them with relevant indicators of weathering degree, the relationship between the two can be established. For example, the relative intensity changes of the absorption characteristics of iron ions and clay minerals in the visible-near infrared spectrum can be used to divide the weathering grade of rocks.

[0043] From the above content, it can be seen that the purpose of this embodiment is to construct a first inversion model through sample data of the sample area, and then substitute the hyperspectral image into the first inversion model to invert the crystallinity, alteration component concentration and weathering degree of the entire area to be predicted.

[0044] The existing technology often only obtains spectral images and directly uses the ready-made inversion model to invert and then obtain the mining area prediction results. That is to say, for the sake of convenience, the existing technology often does not collect and analyze new sample data. This has the following shortcomings: First, the historical inversion data may not be collected from the area to be predicted but from sample data from other areas. Such a prediction process will definitely make the mining area prediction results inaccurate and the mineralization area prediction accuracy low because it is not applicable to the new area to be predicted. Second, the ready-made inversion model is definitely not as accurate as the inversion constructed using new sample data. Therefore, it is definitely not applicable to the new area to be predicted, which will make the mining area prediction results inaccurate and the mineralization area prediction accuracy low.

[0045] The new sample data used in this embodiment for inversion model construction and mining area prediction has accurate data and high data precision, so that the final mining area prediction result has higher precision.

[0046] Step 130: Fusing the fracture features extracted from the high spatial resolution image with the spectral index of the target geological feature factor to obtain a feature matrix.

[0047] It is understandable that fracture characteristics and the degree of rock fragmentation are related.

[0048] Step 140: Input the characteristic matrix into the second inversion model to determine the target mineralization area of ​​the area to be predicted; the second inversion model is a tectonic stress field inversion model pre-constructed based on a graph convolutional neural network.

[0049] In the prior art, after the crack characteristics are obtained, they are directly used to construct the subsequent tectonic stress field inversion model. The factors considered are relatively simple, so the inversion accuracy of the tectonic stress field inversion model is not high. This embodiment associates the three geological characteristic factors of crystallinity, alteration component concentration, and weathering degree with the crack characteristics, so that the graph neural network can learn the relationship between the crack characteristics and more related factors (crystallinity, alteration component concentration, and weathering degree), so that the data basis of the mining area prediction results is more reliable and the prediction accuracy is higher.

[0050] From the above content, it can be seen that the prospecting method for characterizing tectonic stress based on the degree of rock fragmentation provided by the embodiment of the present invention obtains image data and sample data of the area to be predicted, constructs an inversion model using the sample data, and inverts the area to be predicted to obtain target indices corresponding to the crystallinity, concentration of altered components and degree of weathering, thereby solving the problem of low data reliability caused by using only the single data source of image data in the prior art. Then, the crack characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor are fused to obtain a characteristic matrix. Finally, the characteristic matrix is ​​input into the tectonic stress field inversion model based on the graph convolutional neural network to determine the target mineralization area of ​​the area to be predicted. In this way, the output result of the tectonic stress field inversion model is more accurate, thereby improving the prediction accuracy of the mineralization area in the area to be predicted.

[0051] Optionally, the sample data includes X-ray data and spectral reflectance data obtained based on a short-wave infrared spectrometer; see Figure 3 , constructing a first inversion model based on the sample data in step 120 may specifically include: Step 121: constructing a spectral library based on the spectral reflectance data; Step 122: Calculate the index of the target geological characteristic factor based on the X-ray data; Step 123: extracting a target band whose index correlation with the target geological characteristic factor is greater than or equal to a preset value from the spectral library, and constructing a spectral index of the target geological characteristic factor based on the target band; Step 124: Determine a first inversion model based on the spectral index of the target geological characteristic factor.

[0052] Before performing step 121, the spectral reflectance acquisition process is described as follows: The first is to set the spectral range: use the visible light (VIR) - short wave infrared spectrum (SWIR) range: 400nm-2500nm.

[0053] Analyze the characteristic bands of key minerals: Visible light band (400nm-700nm): analyze mineral color characteristics and signs of surface oxidation.

[0054] Near-infrared band (700nm-1000nm): detect the presence of iron ions and their coordination environment.

[0055] Short-wave infrared band (1000nm-2500nm): Extract mineral characteristic absorption peaks (such as OH-, , Al-OH, Mg-OH groups). OH-group absorption: 1400nm-1450nm (related to water); Group absorption: around 2300nm (related to carbonates); Al-OH, Mg-OH group absorption: 2200nm-2350nm (related to alteration minerals such as muscovite and chlorite); Fe-O group absorption: 1000nm-1300nm (related to iron minerals).

[0056] The second step is to prepare the acquisition tool: the acquisition tool can be a shortwave infrared spectrometer (ASDFieldSpec).

[0057] Before collection, rock samples need to be dried and crushed into uniform particles to reduce surface effects.

[0058] Finally, the following collection process is adopted: 1) Set the integration time and optimize the acquisition conditions to ensure the signal-to-noise ratio. For low reflectivity samples (such as dark rocks and weathered minerals), set a longer time (such as 100ms-500ms); for high reflectivity samples (such as muscovite and quartz), set a shorter time (such as 10ms-50ms).

[0059] 2) Collect multiple points for each sample and take the average value to reduce local errors.

[0060] 3) Use ENVI software to calibrate the spectral data (subtract background noise and standardize the spectral curve).

[0061] After the collection is completed, step 121 is performed, and step 121 may specifically include: Data preprocessing steps: radiometric correction to remove instrumental noise, background subtraction, normalization of spectra using blackbody reference data, smoothing (e.g. Savitzky-Golay filtering) to reduce noise.

[0062] Feature extraction step: Automatically detect the absorption peak position, depth and width. Determine the key characteristic bands of the mineral and verify the characteristic peaks.

[0063] Steps for building a spectral library: Import the processed spectral data into the ENVI spectral library module. Assign labels to each rock sample, such as "fracture zone_severely broken", "altered_chlorite". Use ENVI's built-in spectral matching tools (such as Spectral AngleMapper, SAM) to verify spectral consistency. Finally, save the spectral data by classification (such as .sli, .hdr format) to ensure that the spectral library is structured and extensible.

[0064] 4) Spectral library verification: Use known mineral standard spectra (such as USGS or ENVI built-in library) to compare and verify the spectral library. Supplement typical mineral samples with missing spectral data. Eliminate abnormal spectra to ensure data quality.

[0065] In step 122, in calculating the index of the target geological characteristic factor based on the X-ray data, specifically, the index of each target geological characteristic factor is calculated.

[0066] For example, for crystallinity, the X-ray data is X-ray diffraction data obtained based on an X-ray diffraction instrument; step 122 may specifically include: Step 1: Determine the X-ray diffraction pattern based on the X-ray diffraction data; Specifically, the crystal structure of the collected samples is analyzed by X-ray diffraction (XRD) instrument. The characteristic diffraction peaks of the minerals are recorded, and the main mineral types and their distribution characteristics are identified by combining parameters such as peak position, intensity and half-maximum width (FWHM), and finally an X-ray diffraction spectrum is generated.

[0067] Step 2: Based on the X-ray diffraction pattern, use the formula: (1) The crystallinity index is calculated; wherein, is the index of the crystallinity; is the area under a specific diffraction peak; is the total diffraction signal area; is the background signal; 1 is the starting angle; 2 is the end angle; is the incident angle; is the diffraction angle; To calculate the lower limit of integration of the entire diffraction pattern; is the upper limit of the integral for calculating the entire diffraction pattern.

[0068] Regarding the alteration component concentration and weathering degree, step 122 may specifically include: calculating the alteration component concentration index and the weathering degree index based on the X-ray fluorescence data, wherein the X-ray fluorescence data is data obtained based on an X-ray fluorescence instrument.

[0069] For example, for the alteration component concentration, step 122 may specifically include: First, analyzing and obtaining target elements of alteration components in rock samples; the alteration components include at least one or more of chlorite, muscovite and pyrite; the target elements include one or more of Fe, Mg, Al and K; In this embodiment, an XRF instrument is used to perform full element scanning analysis on rock samples and non-mineral samples to determine their main chemical components (such as , , , The mineralization potential and alteration degree of the sample can be preliminarily determined by the element distribution characteristics.

[0070] By comparing the chemical composition of ore samples and surrounding rock samples, the enrichment of corresponding elements in altered minerals (such as chlorite, muscovite, pyrite, etc.) is analyzed in detail.

[0071] Second, extract the concentration of the target element from the X-ray fluorescence data; Third, use the formula: (2) The index of alteration component concentration is calculated; where: is an index of the concentration of alteration components, is the concentration of the target element in the ore sample; is the background concentration of the target element in the surrounding rock sample.

[0072] For example, with respect to the degree of weathering, step 122 may specifically include: Extracting the content of target oxides from X-ray fluorescence data; target oxides include easily weathered oxides and stable oxides; The ratio of the content of easily weathered oxides to the content of stable oxides was determined as an index of the weathering degree.

[0073] In the first specific implementation, the calculation of the weathering degree index is exemplarily described as follows: Based on the XRF measurement of the main element oxide content (such as , , , , , etc.), calculate the chemical weathering index (WPI, Weathering Potential Index) (3) : Represents the total amount of easily weathered alkali metal oxides.

[0074] : Indicates the total amount of stable components and other non-weathering oxides.

[0075] Index range: >1: fresh rock (low weathering degree); <0.5 indicates high weathering (significantly weathered rock or soil).

[0076] As for step 123: extracting a target band whose index correlation with the target geological characteristic factor is greater than or equal to a preset value from the spectral library, and constructing a spectral index of the target geological characteristic factor based on the target band, the following specific steps are included: The first sub-step: using correlation analysis, calculate the correlation coefficient between the reflectivity of each band and the index of the target geological characteristic factor to determine the target band corresponding to each target geological characteristic factor, where the target band refers to the band whose correlation is greater than or equal to a preset value.

[0077] The determined target bands are described as follows: Target bands for crystallinity: 2200nm (Al-OH group) and 2350nm (Mg-OH group).

[0078] Target bands of alteration mineral concentration index: Al-OH-MI (aluminum-hydroxy mineral index) is used to reflect hydroxy minerals (such as muscovite, kaolinite, etc., select 2200nm and 2160nm. FeI-SWIR (iron oxide index) is used to reflect the enrichment of iron-stained minerals (such as limonite, hematite), select 2200nm, 1660nm, 850nm and 650nm.

[0079] Target bands for weathering degree: alkali metal oxides ( , ) related bands Short-wave infrared has weak iron absorption and reflectance characteristics related to easily weathered minerals at 1000nm, while 1750nm reflects the absorption characteristics of alkali metal minerals. These bands reflect the characteristics of hydration and weakly bound alkali metals in rocks. Silicate minerals (such as , )’s absorption characteristic band is not affected by water vapor absorption, so 2200nm and 2350nm are selected. 2200nm reflects the Al-OH absorption characteristics of stable minerals such as muscovite and kaolinite, while 2350nm reflects the Mg-OH absorption characteristics of stable minerals such as chlorite.

[0080] The second sub-step: Use feature extraction algorithm (principal component analysis PCA) to compress spectral dimensions, use PCA to screen out the bands that are most sensitive to the target geological characteristic factors, reduce the interference of non-related bands, and improve the analysis efficiency of spectral data.

[0081] According to the screening bands, the spectral index of crystallinity (CI-SWIR), the spectral index of alteration component concentration (ACI-SWIR) and the spectral index of weathering degree (WPI-SWIR) are constructed, so that there is a clear physical meaning between the spectral index and the geological attributes.

[0082] The specific process of the second sub-step is described in detail as follows: 1) Spectral index of crystallinity (CI-SWIR) As partially explained above, the spectral index of crystallinity is mainly based on the 2200nm and 2350nm bands to construct the crystallinity index of cataclase, and its theoretical concept is as follows: Absorption depth: Absorption depth reflects the absorption intensity of a mineral to light of a specific wavelength. In the mineral spectrum, it is the depth of the lowest reflectivity point (absorption center) of a specific band relative to the spectral baseline.

[0083] Characteristics of high crystallinity: deeper absorption peak (high absorption depth), sharper absorption band.

[0084] The relationship between absorption depth and crystallinity is mainly reflected in how the spectral absorption characteristics of minerals reflect the order of the crystal structure. The absorption depth is positively correlated with the crystallinity of the mineral, and the correlation has the following characteristics: The order of the crystal structure has an enhanced characteristic for spectral absorption: In highly crystalline minerals, the atoms are arranged in an orderly and stable manner, and the absorption band intensity of specific chemical groups (such as Al-OH, Mg-OH) is higher. The ordered structure makes the interaction between light and the atoms inside the mineral stronger, resulting in a deeper absorption band.

[0085] Amorphous characteristics of low-crystallinity minerals: The atoms of low-crystallinity or amorphous minerals are disordered, the absorption of light of specific wavelengths is weakened, the absorption band is shallower and the shape is wider.

[0086] The distribution area of ​​cataclasites often experiences strong hydrothermal alteration, and the absorption peaks of 2200nm and 2350nm often exist at the same time. 2200nm reflects the development of muscovite or kaolinite, and 2350nm reflects the development of chlorite or other magnesium minerals. Al-OH and Mg-OH groups may exist at the same time in cataclasites. In order to more comprehensively characterize cataclasites, the absorption characteristics of 2200nm and 2350nm can be used simultaneously, combining the characteristics of the two through an index.

[0087] The absorption characteristics of 2200nm (Al-OH group) are usually produced by muscovite, kaolinite and other minerals containing Al-OH. This absorption peak is related to the vibration of the Al-OH group in the mineral. Muscovite or kaolinite in cataclasite is often formed by hydrothermal transformation or alteration under tectonic stress, which can characterize the following characteristics of cataclasite: First, the intensity of alteration, high absorption depth usually corresponds to stronger alteration. Second, the change in crystallinity, the displacement and depth of the absorption peak can reflect the degree of order of the crystal structure of the Al-OH group. It is suitable for the characterization of cataclasite dominated by muscovite and kaolinite, especially in the alteration zone.

[0088] The absorption characteristics of 2350nm (Mg-OH group) are produced by chlorite, serpentine and other Mg-OH-containing minerals. This absorption peak is related to the vibration of the Mg-OH group. Mg-containing minerals in cataclasites (such as chlorite and serpentine) are usually associated with alteration and tectonic stress transformation. The following characteristics of cataclasites can be characterized: ① Hydrothermal alteration, the depth and shape of the absorption peak of the Mg-OH group reflect the intensity of hydrothermal activity. ② Changes in mineral composition, the compositional evolution of minerals under tectonic stress can be identified through the absorption characteristics of Mg-OH. It is more suitable for cataclasites containing chlorite or serpentine, especially in areas with significant hydrothermal transformation.

[0089] Spectral index of crystallinity ( ) is a spectral index based on the short-wave infrared (SWIR) band, which is used to characterize the degree of order of mineral crystals. According to the relevant mineral characteristics such as Al-OH groups and Mg-OH groups, the formula is constructed as follows: = (4) Parameter Description: in: . Depbaseline: Determined by the overall spectral baseline of the fractured rock.

[0090] The absorption depth calculation formula is: (4-1) In formula (4-1): is the wavelength, such as 2200nm, 2350nm; is the actual reflectivity of the target band; is the baseline reflectivity of the target band.

[0091] The calculation formula is as follows: (4-2) In formula (4-2): (4-3) In formula (4-3) is the baseline reflectivity of the 2200nm absorption band. is the starting wavelength, for example, the starting wavelength of the 2200nm absorption band is 2150nm; is the end wavelength, for example, the end wavelength of the 2200nm absorption band is 2250nm; the end wavelength of the 2350nm absorption band is 2400nm. is the wavelength The actual reflectivity, is the wavelength The actual reflectivity.

[0092] 2) Spectral index of alteration component concentration (ACI-SWIR) Calculate the Al-OH-MI spectral index: Al-OH-MI (spectral index of aluminum-hydroxy minerals) is used to reflect the enrichment of hydroxy minerals (such as muscovite, kaolinite, etc.), and its calculation formula is: (5) R2200: Reflectivity in the 2200nm band, reflecting the Al-OH absorption characteristics.

[0093] R2160: Reflectance at 2160nm, used as a baseline reference.

[0094] Calculate the Iron Oxide Index (FeI-SWIR): FeI-SWIR (spectral index of iron oxide) is used to reflect the enrichment of iron-stained minerals (such as limonite and hematite). The calculation formula is: (6) R2200: Reflectivity in the 2200nm band, used to indicate the response of iron-dyed minerals to the shortwave infrared band.

[0095] R1660: 1660nm band reflectance, used for baseline correction.

[0096] R850 and R650: Reflectances in the 850nm and 650nm bands respectively, representing the characteristic absorption of iron-dyed minerals in the visible light range.

[0097] Comprehensive alteration component concentration index (ACI-SWIR): Combining the two sub-indices of Al-OH-MI and FeI-SWIR, the spectral index ACI-SWIR of comprehensive alteration component concentration is constructed: ACI-SWIR=w1*Al-OH-MI+w2*FeI-SWIR (7) w 1、 w2 is the weight factor, and both can be adjusted and optimized according to the mineral combination and alteration characteristics of the area to be predicted.

[0098] 3) Spectral index of weathering degree (WPI-SWIR) Total amount of easily weathered alkali metal oxides: mainly including Na2O and K2O. These components are easily affected by chemical weathering and show absorption characteristics in specific bands in the spectrum.

[0099] Stable components and total amount of oxides that are not easily weathered: mainly include SiO2, Al2O3 and other components. These components are more stable during the weathering process and also have significant characteristics in the spectrum.

[0100] The weathering degree index (WPI) indicates the ratio of easily weathered components to stable components. The higher the value, the greater the weathering potential of the rock.

[0101] (8) For step 124: determining the first inversion model based on the spectral index of the target geological characteristic factor, the spectral index of each of the above target geological characteristic factors is a part of the first inversion model, that is, the spectral index of all the target geological characteristic factors constitute the first inversion model. The function of the first inversion model is to invert the hyperspectral image of the area to be predicted collected by the hyperspectral camera to obtain the mineral composition and chemical composition of the area to be predicted (for example, the target index of crystallinity, the target index of alteration component concentration and the target index of weathering degree all reveal the content of their corresponding mineral components and chemical components).

[0102] For example, when using drones to collect high-spatial-resolution images, the drones are equipped with high-definition cameras and follow a preset flight path to cover the area to be predicted. The flight altitude is controlled at 30-50 meters, and the image overlap rate is maintained at more than 80%. The data is used to extract the spatial distribution of crack features, including the direction, length, and density of linear features.

[0103] For example, when using a drone to collect hyperspectral images, a hyperspectral camera carried by the drone is used to capture spectral data in the range of 400-2500nm. The data is used for mineral inversion, including crystallinity, concentration, and weathering degree of altered minerals.

[0104] High spatial resolution images are geometrically corrected and seamlessly mosaicked to generate complete crack distribution images.

[0105] Hyperspectral images use the FLAASH model or QUAC model to eliminate atmospheric effects and use geographic coordinates to align images to ensure consistency between hyperspectral image data and other data sources.

[0106] After obtaining high spatial resolution images, hyperspectral images, crack images, X-ray diffraction data, and X-ray fluorescence data, the multi-source data can be cleaned and standardized, for example, using the standardization method: Min-Max normalization.

[0107] Exemplarily, the process of extracting crack features includes: The high spatial resolution image is processed using a preset deformable convolution and a preset directional filter to obtain a crack feature map; The crack characteristics are extracted from the crack characteristic map; the crack characteristics include crack density, crack direction and crack length.

[0108] Specifically, the preset deformable convolution can enhance the adaptability of the deep learning network model to complex morphological cracks, and the preset directional filter can be combined to highlight the linear crack characteristics.

[0109] The core concept of deformable convolution: Standard convolution uses a fixed-size sampling window in space (such as 3×3). Deformable convolution introduces an offset so that the convolution kernel can dynamically adjust the sampling position to better adapt to the irregular shape of the crack.

[0110] Convolutional feature extraction: The initial convolution layer in the deformable convolution is used to extract the edge, texture and grayscale features of the high spatial resolution image. The deformable convolution layer dynamically adjusts the position of the sampling window to enhance the adaptability to cracks.

[0111] Loss Function: Optimize the identification of crack regions using a specific loss for edge detection.

[0112] The design of the directional filter is introduced below.

[0113] Function of the filter: Cracks usually have linear or strip-like features. Directional filters can enhance the linear structure along a specific direction in the image and suppress background interference.

[0114] Exemplarily, a Fourier directional filter is used, and the filter kernel function is as follows; (9) is the filter kernel function; : The direction of the filter; σ: Gaussian kernel width, which controls the scale of the filter; : Center frequency, determines the sensitive frequency of the filter.

[0115] against =0 degrees, 15 degrees, 30 degrees, 45 degrees, 60 degrees, 75 degrees, 90 degrees, 105 degrees, 120 degrees, 135 degrees, 150 degrees, 165 degrees directions, and filtering is performed successively.

[0116] Parameter setting: 0 degrees ; 15 degrees ; 30 degrees ; 45 degrees ; 60 degrees ; 75 degrees ; 90 degrees ; 105 degrees ; 120 degrees ; 135 degrees ; 150 degrees ; 165 degrees ; The filtering results in all directions are combined to generate an enhanced crack feature map.

[0117] After obtaining the crack feature map, crack extraction and post-processing are performed: 1) Binarization processing: (10) is a binary image; This is the crack feature map after directional filtering.

[0118] The filtered feature image is converted into a binary image, and the threshold T is set. The value of T is determined by the Otsu method.

[0119] 2) According to the degree of fragmentation of rock samples, combined with the direction, length and density of cracks, the stress response of different regions is derived on high spatial resolution image data: High fragmentation area: stress concentration, high crack density and diverse directions. Medium fragmentation area: stress is weak, and the crack direction tends to be consistent. Low fragmentation area: stress is the weakest and cracks are sparse.

[0120] The fused data include fracture characteristics (length, direction, density) and mineral spectral characteristics (crystallinity, concentration of alteration components and degree of weathering, etc.). The above multi-source data generate a unified feature matrix to construct the sample space.

[0121] Feature matrix form: (11) : The first target index to the nth target index of crystallinity. : The first target index to the nth target index of alteration component concentration. : The first target index to the nth target index of weathering degree. : The length of the first crack to the length of the nth crack. : The first crack density to the nth crack density. : The direction of the first crack to the direction of the nth crack.

[0122] Fracture characteristics (length, direction, density) reveal the degree of rock fragmentation, while mineral crystallinity and concentration provide information on the response of the stress field to the distribution of rock minerals. The degree of weathering reflects the chemical stability of the rock, further improving the understanding of the stress field. The target indexes of fracture characteristics, mineral crystallinity, alteration component concentration, and weathering degree are fused. Based on the fusion feature matrix, the sample data of the sample area and the image data of the area to be predicted are used to invert the tectonic stress field of the area to be predicted and predict the ore-rich area of ​​the area to be predicted. Fracture characteristics, mineral distribution characteristics, and weathering index are finally unified into a feature matrix, which can establish a quantitative relationship between the tectonic stress field and multiple variables. Data fusion not only improves the prediction accuracy of the model, but also enhances the interpretation of the results for geological phenomena.

[0123] After obtaining the above-mentioned feature matrix, a part of the feature matrix is ​​used as a training set and the other part as a test set to train and verify the pre-constructed graph convolutional neural network to obtain the second inversion model, namely the tectonic stress field inversion model.

[0124] The pre-built graph convolutional neural network can learn the relationship between stress field and cracks and mineral distribution. It can automatically extract nonlinear features in high-dimensional data and is suitable for complex multi-source data fusion analysis.

[0125] The tectonic stress field inversion model obtained after training and verification can simulate the stress field distribution in combination with geological spatial relationships. Simulating geological spatial relationships (such as the spatial correlation of fracture distribution and mineral concentration) can more accurately predict the spatial variation of the stress field.

[0126] When using the training set to train the graph convolutional neural network, first input the fusion feature matrix (crack features + mineral features + weathering index). Then train the graph convolutional neural network to minimize the error between the predicted stress field and the actual stress distribution. The training process is as follows: Batch Size: Set to 8.

[0127] Optimization algorithm: Synchronous stochastic gradient descent (SGD) is used with weight decay set to 0.0001 and momentum set to 0.9.

[0128] Loss function: Binary Cross-Entropy Loss is used to optimize the crack segmentation results.

[0129] Iteration number: Train the model until the loss of the validation set is stable or minimized.

[0130] After obtaining the tectonic stress field inversion model, data preprocessing, model cleaning and input data standardization are performed; fracture extraction and the first inversion model are processed in parallel; the data fusion model generates a unified feature matrix; and the tectonic stress field inversion model outputs the final result.

[0131] After obtaining the final results of the stress field inversion model output, mineralization potential mapping is carried out to finely divide the mineralization potential area and calibrate the prospecting direction.

[0132] Optionally, the prospecting target area can be optimized based on comprehensive indicators such as stress field, fracture characteristics, mineral concentration and weathering potential.

[0133] When mapping the mineralization potential, GIS tools are used to generate mineralization potential mapping units based on the stress field inversion results and mineralization potential indicators, and the mineralization potential is evaluated in stages: High potential area: The stress concentration in the mapping unit is high, the fracture density is large, the mineral concentration is significant, the alteration component is enriched, the crystallinity is high, the weathering potential index shows strong stability, and it presents strong mineralization indications.

[0134] Medium potential area: The stress characteristics within the mapping unit are moderate, the fracture direction is relatively consistent, the mineral concentration and alteration components are at a medium level, the weathering potential index is moderate, and it has a certain mineralization possibility.

[0135] Low potential area: The stress field effect within the mapping unit is not significant, the cracks are sparse, the mineral concentration is low, the weathering potential index is high, and there are no significant signs of mineralization.

[0136] Through mineralization potential mapping and zoning, high potential areas are clearly marked to form the preliminary scope of the target mineralization area, that is, the prospecting target area. Figure 4 .

[0137] For example, after determining the target mineralization area, the prospecting target area can be optimized by combining the historical mineralization belt and tectonic background, and the mineralization potential mapping results are superimposed and analyzed and verified with the historical mineralization belt data, the distribution of the main fault zones and the geological structure background in the area. See the following steps: Structural trend superposition analysis: Based on the trends of major fault zones and fracture zones, the correlation between high potential mapping units and fault systems is verified.

[0138] Verification of historical mineralized zones: Analyze whether high-potential mapping units coincide with known mineralized areas as a basis for model verification.

[0139] Comprehensive analysis of geological background: Combined with the lithology, tectonic intensity and distribution characteristics of altered minerals, the boundaries of the prospecting target area are further optimized.

[0140] For example, after the target mineralization area is determined, the results can be visualized and a mineralization potential distribution map can be generated, as shown in the following steps: Generate intuitive mineralization potential map: Based on the mineralization potential zoning, use GIS tools to generate mineralization potential distribution map, marking the scope and priority of high potential target areas.

[0141] Integrated multi-dimensional data display: The mineralization potential map combines information such as fault zones, fracture zones, historical deposits, and geological structures to form a multi-dimensional superimposed prospecting auxiliary decision map. The significance of each sub-area includes: Prioritize detailed geological exploration and sample collection in high-potential areas. Conduct additional surveys in medium-potential areas to verify the reliability of model predictions. Reduce resource investment in low-potential areas and retain them only as secondary target areas.

[0142] From the above content, it can be seen that the distribution of tectonic stress field and ore body enrichment is not only affected by the characteristics of cracks, but also related to mineral distribution and weathering degree. A single data source is difficult to fully reflect the complexity of the stress field. Therefore, the method of using a single data source to find the target mineralization area in the prior art is not accurate enough, and the obtained mineral distribution map is not accurate enough. This embodiment uses short-wave infrared spectrometer (SWIR), X-ray diffractometer (XRD) and X-ray fluorescence spectrometer (XRF) to collect multi-source data such as spectral characteristics, mineral crystallinity and chemical composition of rock samples in the sample area, extract the crystallinity, alteration component concentration, weathering degree and crack characteristics (length, direction and density) of the mineral; use a drone equipped with a high-definition camera and a hyperspectral camera, combine the collected data with preprocessing and feature extraction, and generate a unified multimodal feature matrix; use a graph convolutional neural network (GCN) to construct a stress field inversion model to simulate the distribution of cracks and stress field characteristics; according to the model output, generate a mineralization potential distribution map of the study area, delineate high potential areas, and optimize prospecting targets. The present invention realizes for the first time an integrated technical solution from rock fragmentation degree mapping to tectonic stress field inversion and ore body enrichment area prediction, which is efficient, accurate and applicable, and is particularly suitable for mineral exploration under complex geological backgrounds.

[0143] like Figure 5 As shown, the present invention also provides a prospecting system for characterizing tectonic stress based on rock fragmentation degree, including: a data acquisition module 510 and a data processor 520; the data acquisition module 510 is in communication connection with the data processor 520; The data acquisition module 510 includes a drone; the drone is equipped with a high-definition camera and a hyperspectral camera; the high-definition camera is used to collect high spatial resolution images of the area to be predicted, and the hyperspectral camera is used to collect hyperspectral images of the area to be predicted; Data processors are used to: Obtaining image data and sample data of the area to be predicted, wherein the image data includes a hyperspectral image and a high spatial resolution image, and the sample data is data corresponding to rock samples in a sample area in the area to be predicted; A first inversion model is constructed based on the sample data, and the area to be predicted is inverted based on the first inversion model and the hyperspectral image to obtain the spectral index of the target geological characteristic factors of the area to be predicted; the target geological characteristic factors include crystallinity, alteration component concentration and weathering degree; The crack characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor are fused to obtain a characteristic matrix; The characteristic matrix is ​​input into the second inversion model to determine the target mineralization area in the area to be predicted; the second inversion model is a tectonic stress field inversion model pre-constructed based on a graph convolutional neural network.

[0144] Optionally, the data acquisition module 510 may also include a short-wave infrared spectrometer, an X-ray diffraction instrument, and an X-ray fluorescence instrument; The short-wave infrared spectrometer is used to analyze and obtain the spectral reflectance data of the rock samples in the sample area; X-ray diffraction instruments are used to collect X-ray diffraction data of rock samples; X-ray fluorescence instruments are used to collect X-ray fluorescence data of rock samples.

[0145] See also Figure 6 For example, a prospecting system for characterizing tectonic stress based on rock fragmentation degree may include: an integrated multimodal data collector, a data processor, and a visualization and operation platform. The integrated multimodal data collector is linked to the data processor in communication, and the data processor is linked to the visualization and operation platform in communication.

[0146] The integrated multimodal data collector is used to collect multi-source data from the mining environment to provide basic information for subsequent analysis and modeling. The integrated multimodal data collector is specifically used for multi-source data collection: collecting fracture distribution information, mineral characteristics, spectral characteristics, chemical composition of geological rock samples, and realizing automated and portable operations, which is suitable for complex field environments.

[0147] The integrated multimodal data collector may specifically include: The rock fragmentation degree recorder takes high-resolution images of the mining area, extracts the length, direction and density characteristics of the cracks, and automatically compares the degree of rock fragmentation in the sample area.

[0148] UAV system: equipped with high-definition camera and hyperspectral camera. High-definition camera is used to achieve high spatial resolution data collection. Configure GPS and flight control system to ensure image coverage and data accuracy. Hyperspectral camera: carried on the UAV, collects hyperspectral images in the range of 400-2500nm, and realizes the rapid collection of spectral characteristic data over a large range of mining areas.

[0149] Short-wave infrared spectrometer (SWIR): measures the short-wave infrared spectrum of rock samples and extracts parameters such as absorption depth, peak position and peak area.

[0150] XRD analysis equipment (X-ray diffractometer): Determines the mineral crystallinity of rock samples and provides crystal structure information. Resolves the characteristic diffraction peak position, half-width (FWHM) and intensity of minerals.

[0151] XRF spectrometer (X-ray fluorescence spectrometer): quickly determines the chemical composition of rock samples (such as oxide content) and assesses alteration components and weathering potential.

[0152] The data processor is used to clean, process, fuse and model the collected multi-source data, and finally output the stress field distribution and ore body prediction results.

[0153] Data processors include: High-performance computing unit: equipped with GPU acceleration module for training and prediction of deep learning models. Supports high parallel computing capabilities, processing multi-source data fusion and large-scale stress field simulation.

[0154] The ENVI software in the analysis software tools is used for radiation correction, band screening and feature extraction of hyperspectral data. The ArcGIS software in the analysis software tools is used for data visualization and spatial analysis.

[0155] Deep Learning Framework: Use TensorFlow or PyTorch to build a graph convolutional neural network (GCN) model.

[0156] Data processing unit: The standardization tool is used to unify the dimensions of multi-source data. The feature extraction module is used to extract key feature parameters from spectral data, image data and chemical composition.

[0157] Model training and prediction unit: Apply deep learning algorithms to perform stress field inversion. Based on fracture characteristics and mineral properties, calibrate mineralization potential areas.

[0158] The visualization and operation platform is used to provide result display and user operation interface, to present prediction results and assist mining exploration decision-making, as well as to visualize tectonic stress field and mineralization potential regional distribution map. It provides intuitive mining exploration data support and assists field operation planning.

[0159] The visualization and operation platform specifically includes the following components: GIS system: used to generate tectonic stress field distribution maps and mineralization potential area prediction maps.

[0160] Result overlay tool: used to combine fault zones, broken zones, and mineral feature distribution to generate multi-dimensional prediction maps.

[0161] User operation terminal: used for user parameter setting, etc. Portable display device in the user operation terminal: with touch function, used to view analysis results in real time in the field. Interactive interface in the user operation terminal: supports operators to select input data range, adjust analysis parameters, and generate prediction maps in real time.

[0162] Cloud storage and remote collaboration: Cloud servers can store analysis results and support remote data access. The collaborative platform allows multiple users to share real-time data and improve mining exploration efficiency.

[0163] Although the present invention is described herein in conjunction with various embodiments, in the process of implementing the claimed invention, those skilled in the art may understand and implement other variations of the disclosed embodiments by viewing the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other components or steps, and "one" or "an" does not exclude multiple situations. A single processor or other unit may implement several functions listed in the claims. Certain measures are recorded in different dependent claims, but this does not mean that these measures cannot be combined to produce good results.

[0164] Although the present invention has been described in conjunction with specific features and embodiments thereof, it is apparent that various modifications and combinations may be made thereto without departing from the spirit and scope of the present invention. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present invention as defined by the appended claims and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present invention. Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, the present invention is intended to include such modifications and variations if they fall within the scope of the claims of the present invention and their equivalents.

Claims

1. A prospecting method based on characterizing tectonic stress by the degree of rock fragmentation, characterized in that: include: Acquire image data and sample data of the area to be predicted, wherein the image data includes a hyperspectral image and a high spatial resolution image, and the sample data is data corresponding to rock samples in a sample area in the area to be predicted; A first inversion model is constructed based on the sample data, and the area to be predicted is inverted based on the first inversion model and the hyperspectral image to obtain a spectral index of a target geological characteristic factor of the area to be predicted; the target geological characteristic factor includes crystallinity, alteration component concentration and weathering degree; Performing data fusion on the fracture characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor to obtain a characteristic matrix; The characteristic matrix is ​​input into a second inversion model to determine the target mineralization area of ​​the area to be predicted; the second inversion model is a tectonic stress field inversion model pre-constructed based on a graph convolutional neural network.

2. The method for prospecting based on characterizing tectonic stress by the degree of rock fragmentation according to claim 1, characterized in that: The sample data includes spectral reflectance data and X-ray data obtained based on a short-wave infrared spectrometer; Constructing a first inversion model based on the sample data includes: Based on the spectral reflectance data, construct a spectral library; Calculating the index of the target geological characteristic factor based on the X-ray data; Extracting a target band whose exponential correlation with the target geological characteristic factor is greater than or equal to a preset value from the spectral library, and constructing a spectral index of the target geological characteristic factor based on the target band; The first inversion model is determined based on the spectral index of the target geological characteristic factor.

3. The method for prospecting based on characterizing tectonic stress by rock fragmentation degree according to claim 2, characterized in that: The X-ray data at least includes X-ray diffraction data obtained based on an X-ray diffraction instrument; The step of calculating the index of the target geological characteristic factor based on the X-ray data comprises: Determining an X-ray diffraction pattern based on the X-ray diffraction data; Based on the X-ray diffraction pattern, the formula is used: ； The crystallinity index is calculated; wherein, is the index of the crystallinity; is the area under a specific diffraction peak; is the total diffraction signal area; is the background signal; 1 is the starting angle; 2 is the end angle; is the incident angle; is the diffraction angle; is the lower limit of the integral; The upper limit of points.

4. The method for prospecting based on characterizing tectonic stress by rock fragmentation degree according to claim 3, characterized in that: The X-ray data also includes X-ray fluorescence data obtained based on an X-ray fluorescence instrument; The step of calculating the index of the target geological characteristic factor based on the X-ray data comprises: Based on the X-ray fluorescence data, an index of the alteration component concentration and an index of the weathering degree are calculated.

5. The method for prospecting based on characterizing tectonic stress by rock fragmentation degree according to claim 4, characterized in that: Determining the index of the alteration component concentration based on the X-ray fluorescence data comprises: Analyze and obtain target elements of alteration components in the rock sample; the alteration components include at least one or more of chlorite, muscovite and pyrite; the target elements include one or more of Fe, Mg, Al and K; Extracting the concentration of the target element from the X-ray fluorescence data; Using the formula: ； The index of the concentration of the alteration component is calculated; wherein, is an index of the concentration of the alteration component, is the concentration of the target element; is the background concentration of the target element.

6. The method for prospecting based on characterizing tectonic stress by the degree of rock fragmentation according to claim 5, characterized in that: Determining the weathering degree index based on the X-ray fluorescence data includes: Extracting the content of target oxides from the X-ray fluorescence data; the target oxides include easily weathered oxides and stable oxides; The ratio of the content of the easily weathered oxide to the content of the stable oxide is determined as the index of the weathering degree.

7. The method for prospecting based on characterizing tectonic stress by rock fragmentation degree according to claim 1, characterized in that: The process of extracting the crack characteristics includes: Processing the high spatial resolution image using a preset deformable convolution and a preset directional filter to obtain a crack feature map; The crack characteristics are extracted from the crack characteristic map; the crack characteristics include crack density, crack direction and crack length.

8. The method for prospecting based on characterizing tectonic stress by rock fragmentation degree according to claim 1, characterized in that: The step of inputting the characteristic matrix into the second inversion model to determine the target mineralization area of ​​the area to be predicted comprises: Inputting the characteristic matrix into the second inversion model to obtain the tectonic stress field distribution of the area to be predicted; Based on the distribution of the tectonic stress field and the preset mapping benchmark, a mineralization potential mapping unit is generated using a GIS tool; the mineralization potential mapping unit includes a high potential area, a medium potential area and a low potential area; The high potential area is determined as the target mineralization area.

9. A prospecting system based on characterizing tectonic stress by the degree of rock fragmentation, characterized in that: include: Data acquisition module and data processor; The data acquisition module is in communication with the data processor; The data acquisition module includes a drone; the drone is equipped with a high-definition camera and a hyperspectral camera; the high-definition camera is used to collect high-spatial-resolution images of the area to be predicted, and the hyperspectral camera is used to collect hyperspectral images of the area to be predicted; The data processor is used to obtain image data and sample data of the area to be predicted, wherein the image data includes a hyperspectral image and a high spatial resolution image, and the sample data is data corresponding to rock samples in a sample area in the area to be predicted; A first inversion model is constructed based on the sample data, and the area to be predicted is inverted based on the first inversion model and the hyperspectral image to obtain a spectral index of a target geological characteristic factor of the area to be predicted; the target geological characteristic factor includes crystallinity, alteration component concentration and weathering degree; Performing data fusion on the fracture characteristics extracted from the high spatial resolution image and the spectral index of the target geological characteristic factor to obtain a characteristic matrix; The characteristic matrix is ​​input into a second inversion model to determine the target mineralization area of ​​the area to be predicted; the second inversion model is a tectonic stress field inversion model pre-constructed based on a graph convolutional neural network.

10. The mineral prospecting system based on characterizing tectonic stress by rock fragmentation degree according to claim 9, characterized in that: The data acquisition module also includes a short-wave infrared spectrometer, an X-ray diffraction instrument and an X-ray fluorescence instrument; The short-wave infrared spectrometer is used to analyze and obtain high-spectral data of rock samples in the sample area; The X-ray diffraction instrument is used to collect X-ray diffraction data of the rock sample; The X-ray fluorescence instrument is used to collect X-ray fluorescence data of the rock sample.

Images