Core mineral category identification method and system, computer storage medium

Abstract

Embodiments of the present application relate to the field of analyzing materials by optical means, and in particular to a method and system for identifying mineral categories of a core, and a computer storage medium. The method comprises the following steps: S10: obtaining a core of a deep borehole; S20: obtaining hyperspectral image data and high-resolution image data of the core, and obtaining hyperspectral high-resolution fusion image data of the core; S30: segmenting the image data obtained in step S20; S40: determining a representative spectrum of each part of the segmented image data; S50: determining a mineral category of each part according to the representative spectrum; and S60: determining a mineral category of the core according to the mineral category of each part. The method and system improve data quality and accuracy by obtaining hyperspectral high-resolution fusion image data of the core, mitigate spectral mixing effects, and reduce misjudgments; by segmenting the fusion image data and generating a representative spectrum, pixel-by-pixel processing can be replaced to reduce computational complexity and improve identification efficiency.

Description

Technical Field

[0001] The embodiments of this application relate to the field of material analysis using optical means, and particularly to a method and system for identifying mineral types in rock cores, and a computer storage medium. Background Technology

[0002] The statements herein are provided merely as background information in connection with this application and do not necessarily constitute prior art.

[0003] Deep borehole cores are rock samples obtained from deep boreholes. They can directly reflect the original state of the strata. Analyzing the mineral types of the cores helps to analyze the lithology of the strata, reservoir properties, and structural information.

[0004] Traditional core analysis methods require core sampling and slide preparation, which may damage the core samples and prevent repeated testing of the same sample. This results in limited accuracy, low efficiency, and incomplete information. In contrast, spectral scanning technology enables non-destructive, continuous, and comprehensive digital acquisition of core samples, obtaining more accurate spectral and textural information, leading to more objective and standardized mineral identification. Summary of the Invention

[0005] A brief overview of this application is provided below to offer a basic understanding of certain aspects thereof. It should be understood that this overview is not an exhaustive summary of the application. It is not intended to identify key or essential parts of the application, nor is it intended to limit its scope. Its purpose is merely to present certain concepts in a simplified form as a prelude to the more detailed description that follows.

[0006] This application provides a method for identifying the mineral category of a rock core, comprising the following steps: S10: acquiring a rock core from a deep borehole; S20: acquiring hyperspectral image data and high-resolution image data of the rock core, processing the data to improve the spatial resolution of the hyperspectral image data, and obtaining fused hyperspectral high-resolution image data of the rock core; S30: segmenting the image data obtained in step S20; S40: determining the representative spectrum of each part of the segmented image data; S50: determining the mineral category of each part based on the representative spectrum; S60: determining the mineral category of the rock core based on the mineral category of each part.

[0007] Compared to the pixel spectral mixing problem commonly found in existing identification methods and the limitations of using only simple image processing or radiometric correction, the mineral classification identification method for rock cores provided in this application improves the quality and accuracy of source data for mineral identification by obtaining high-spectral, high-resolution fused image data of the rock core. This alleviates the problem of feature band weakening and shifting caused by the mixing of multiple minerals within a pixel, enhances the distinguishability of mineral absorption bands, and thus more accurately distinguishes different minerals with similar spectral characteristics, reducing misjudgments. By segmenting the fused image data and generating representative spectra, pixel-by-pixel processing can be replaced to reduce computational load and improve identification efficiency.

[0008] This application, in another aspect, provides a mineral classification system for rock cores, comprising: a sensor configured to acquire rock cores from deep boreholes; and a processor configured to: acquire hyperspectral image data and high-resolution image data of the rock cores; process the data to improve the spatial resolution of the hyperspectral image data to obtain fused hyperspectral high-resolution image data of the rock cores; segment the fused image data; determine a representative spectrum for each segment of the segmented image data; determine the mineral classification of each segment based on the representative spectrum; and determine the mineral classification of the rock core based on the mineral classification of each segment.

[0009] The mineral classification identification system for rock cores provided in this application improves the quality and accuracy of source data for mineral identification by obtaining high-spectral, high-resolution fused image data of the rock cores. This alleviates the problem of weakening and shifting of characteristic spectral bands caused by the mixing of multiple minerals within a pixel, enhances the distinguishability of mineral absorption bands, and thus more accurately distinguishes different minerals with similar spectral characteristics, reducing misjudgments. By segmenting the fused image data and generating representative spectra, it can replace pixel-by-pixel processing to reduce computational load and improve identification efficiency.

[0010] This application also provides a computer storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, perform the aforementioned method. Attached Figure Description

[0011] To further illustrate the above and other advantages and features of this application, the specific embodiments of this application will be described in more detail below with reference to the accompanying drawings. The drawings, together with the following detailed description, are included in and form a part of this specification. Elements having the same function and structure are indicated by the same reference numerals. It should be understood that these drawings only depict typical examples of this application and should not be considered as limiting the scope of this application.

[0012] Figure 1 This is a flowchart illustrating a method for identifying mineral types in rock cores according to an embodiment of this application; Figure 2 This is a schematic diagram comparing the physical morphology of hyperspectral image data and high-resolution image data before and after fusion processing in the method provided in the embodiments of this application; Figure 3 This is a schematic diagram showing the distribution of altered minerals at different depths, obtained by using the identification method provided in the embodiments of this application. Figure 4 This is a schematic diagram showing the distribution of alteration minerals in a favorable mineralization zone, obtained by using the identification method provided in the embodiments of this application.

[0013] Explanation of reference numerals in the attached figures: 1. High-resolution images; 2. Visible-near-infrared hyperspectral images; 3. Short-wave infrared hyperspectral images; 4. Hyperspectral high-resolution fused images. Detailed Implementation

[0014] Exemplary embodiments of this application will be described below with reference to the accompanying drawings. For clarity and brevity, not all features of actual implementations are described in the specification. However, it should be understood that many implementation-specific decisions must be made in the development of any such actual embodiment to achieve the developer's specific goals, such as complying with constraints related to the system and business, and these constraints may vary depending on the implementation. Furthermore, it should be understood that while development work can be very complex and time-consuming, such development work is merely a routine task for those skilled in the art who benefit from the content of this application.

[0015] It should also be noted that, in order to avoid obscuring this application with unnecessary details, only the equipment structure and / or processing steps closely related to the solution according to this application are shown in the accompanying drawings, while other details that are not closely related to this application are omitted.

[0016] The following disclosure provides several different implementations or examples for carrying out this application. To simplify the disclosure of this application, specific examples of components and methods are described below. Of course, these are merely examples and are not intended to limit this application. In the description of the embodiments of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0017] In current core mineral classification applications, the use of spectral scanning technology often suffers from pixel spectral aliasing. Since a single pixel often covers multiple minerals, diagnostic information such as the characteristic absorption bands of key minerals, including band position, band depth, and bandwidth, is severely diluted or shifted, making traditional pixel-by-pixel spectral matching methods prone to misclassification. Secondly, the inherent multi-band noise in hyperspectral data and instrument differences, including differences in radiometric response across different bands, sensor strip noise, and variations in ambient light, affect the stability of spectral similarity measurements. Furthermore, the diversity of mineral species and the similarity of spectral features are intertwined. Especially for typical hydrothermal alteration minerals such as chlorite, sericite, and carbonates, whose characteristic absorption bands are located close together, more robust measurement methods and cross-verification with multiple pieces of evidence are needed for effective differentiation. Finally, traditional pixel-by-pixel recognition methods are computationally intensive and redundant, resulting in slow processing speeds that are difficult to meet the needs of online applications. Moreover, in the subsequent statistical generation of regional mineral names, fragmentation errors are easily introduced due to misclassification of boundary pixels.

[0018] In summary, existing deep borehole core mineral identification technologies struggle to achieve an effective balance between accuracy, speed, and automation. There is a lack of a method that can overcome spectral mixing and noise interference while simultaneously enabling rapid, accurate, and automated identification.

[0019] To address the aforementioned problems, embodiments of this application provide a method for identifying the mineral categories of rock cores. Figure 1 This is a flowchart illustrating a method for identifying mineral types in rock cores according to an embodiment of this application, as shown below. Figure 1 As shown, it includes the following steps: S10: Obtaining core samples from deep boreholes; S20: Obtaining hyperspectral image data and high-resolution image data of the core samples, processing the data to improve the spatial resolution of the hyperspectral image data, and obtaining fused hyperspectral high-resolution image data of the core samples; S30: Segmenting the image data obtained in step S20; S40: Determining the representative spectrum of each part of the segmented image data; S50: Determining the mineral category of each part based on the representative spectrum; S60: Determining the mineral category of the core based on the mineral category of each part.

[0020] Compared to the pixel spectral mixing problem commonly found in existing identification methods and the limitations of using only simple image processing or radiometric correction, the mineral classification identification method for rock cores provided in this application improves the quality and accuracy of source data for mineral identification by obtaining high-spectral, high-resolution fused image data of the rock core. This alleviates the problem of feature band weakening and shifting caused by the mixing of multiple minerals within a pixel, enhances the distinguishability of mineral absorption bands, and thus more accurately distinguishes different minerals with similar spectral characteristics, reducing misjudgments. By segmenting the fused image data and generating representative spectra, pixel-by-pixel processing can be replaced to reduce computational load and improve identification efficiency.

[0021] Figure 2 This is a schematic diagram comparing the physical morphology of hyperspectral image data and high-resolution image data before and after fusion processing in the method provided in the embodiments of this application, as shown in the diagram. Figure 2 As shown, high-resolution image 1 can be obtained by a high-resolution digital camera. Before fusion, visible-near-infrared hyperspectral image 2 and short-wave infrared hyperspectral image 3 are limited by the physical limitations of the sensors, resulting in low spatial resolution. This leads to severe blurring and rough mosaic-like appearance of the boundaries of micro-cracks, vein-like minerals, and phenocryst grains on the core surface. However, after fusion, the high-resolution fused hyperspectral image 4 retains the rich spectral dimensional information of the original hyperspectral data while also preserving the physical geometric texture of high-resolution image 1. This makes the micro-geological structure of the core clearly visible, and the originally blurred mineral contact boundaries are sharpened and restored. This indicates that the method provided by the embodiments of this application can effectively extract and align spatial texture features and structural features, realize the lossless injection of spatial detail information, and improve the quality of the source data.

[0022] In traditional low-resolution hyperspectral imagery, a large pixel often covers multiple coexisting minerals. This mixed pixel can cause the absorption bands of characteristic spectra to be diluted or shifted, leading to fragmentation-related misinterpretations, such as... Figure 2 As shown, the method provided in the embodiments of this application improves the spatial resolution of the hyperspectral high-resolution fused image 4 by fusing high-resolution image 1, so that the originally mixed coarse pixels are physically mixed into multiple fine pixel units with independent spectral properties, so that subsequent image segmentation can be strictly carried out along the real mineral physical boundary, thereby fundamentally avoiding the problem of extracting mixed spectra across different mineral grain boundaries, which helps to generate high-purity regional representative spectra and achieve high-precision mineral category identification.

[0023] In some embodiments, step S20 further includes the following steps: S21: Simultaneously extracting and aligning the spatial texture features of the high-resolution image data with the spatial structural features of the hyperspectral image data to obtain first data; S22: Processing the first data based on sparse representation theory to obtain second data, wherein the second data is spatial detail information of joint sparse representation; S23: Injecting the spatial detail information into the hyperspectral image data to obtain a hyperspectral high-resolution image. This enables the main features of the high-resolution image data and the hyperspectral image data to be represented as a sparse combination, maintaining accuracy while reducing redundant information, making the expression more concise.

[0024] In some embodiments, step S21 further includes the following steps: S211: Performing IHS (Intensity-Hue-Saturation Transform) on the hyperspectral image data to obtain the simulated panchromatic component of the hyperspectral image data; S212: Performing histogram matching between the high-resolution image and the simulated panchromatic component to obtain the matched simulated panchromatic component and the high-resolution image; S213: Performing discrete wavelet transform on the simulated panchromatic component to obtain the multi-scale structural details of the simulated panchromatic component; S214: Performing edge-preserving multi-scale texture extraction processing on the high-resolution image to obtain the edge and texture detail information of the high-resolution image; the first data includes the multi-scale structural details of the simulated panchromatic component of the hyperspectral image data and the edge and texture detail information of the high-resolution image. Through multi-scale guided filtering, the output result can have good edge-preserving characteristics, and by considering the multi-scale structural details and edge and texture detail information, relatively complete spatial information can be obtained.

[0025] In some embodiments, step S22 further includes the following steps: S221: Constructing an overcomplete dictionary using the multi-scale structural details of the simulated panchromatic components in the first data and the edge and texture detail information of the high-resolution image in the first data; S222: Determining the sparse representation coefficients of the multi-scale structural details and the edge and texture detail information using the overcomplete dictionary; S223: Fusing the sparse representation coefficients using the L1 norm maximization rule based on region energy weighting (the sum of the absolute values ​​of all elements in the L1 norm representation vector) to obtain the second data. By determining the sparse representation coefficients of the multi-scale structural details and the edge and texture detail information, and then using the L1 norm maximization rule based on region energy weighting for fusion, redundant information injection can be avoided, the influence of redundant information can be filtered out, and the accuracy of spatial detail information can be improved, thereby improving the quality of the fused image.

[0026] In some embodiments, in step S221, an overcomplete dictionary can be constructed using the SVD (Singular Value Decomposition) algorithm, which can reduce the time complexity of the algorithm while ensuring learning efficiency, and at the same time make the constructed overcomplete dictionary more adaptive.

[0027] When determining the input data, a sliding window of a preset size can be used to simultaneously traverse and sample the multi-scale structural details of the simulated panchromatic components and the edge and texture details of the high-resolution image obtained in step S214. Each image patch extracted from the sliding window is used as a typical geological pattern sample of the core at a local location. The extracted image patches are then vectorized and stretched, and the DC component is removed to eliminate the interference of the basic brightness background difference in different regions on feature extraction. All processed vectors are combined and arranged column-wise to construct a joint training sample set matrix as input for dictionary learning, which can reflect all the potential structural commonalities and complementary relationships in the image to be fused.

[0028] The joint training sample set matrix is ​​input into the SVD algorithm. A portion of the training sample data is randomly selected as the initial dictionary to establish the starting point for optimization. Using the initial dictionary, each sample in the joint training sample set matrix is ​​sparsely encoded to obtain a sparse coefficient matrix. Based on the joint training sample set matrix, the sparse coefficient matrix, and the initial dictionary, the updated initial dictionary and the updated sparse coefficient matrix are determined and iterated repeatedly to alternately update the expression of the samples and the dictionary atoms themselves. Each column of atoms in the dictionary is locked and corrected one by one until the optimal solution is found, thereby constructing an overcomplete dictionary. This allows each primitive in the dictionary to evolve step by step, accurately matching specific geological features such as fractures on the core surface and the edges of mineral crystals.

[0029] By constructing a complete dictionary in the above manner, a general feature library containing high-frequency structural features and texture details of rock cores can be established through an unsupervised training process. This captures key geological detail patterns such as surface cracks and mineral crystal edges of rock cores, thereby enabling the construction of a substrate that can accurately map the complex textures and spectral features of deep rock cores. This provides the necessary prior model for accurate sparse decomposition in the subsequent S222 step.

[0030] In some embodiments, the SOMP algorithm can be used to determine the sparse representation coefficients in step S222. For example, based on the overcomplete dictionary constructed in step S221, a mathematical solution based on the synchronous orthogonal matching pursuit strategy (SOMP) is performed on the first data that can characterize the core features to calculate the precise expression values ​​of hyperspectral image features and high-resolution image features in the same sparse domain, thus transforming the image features in the physical space into sparse representation coefficients in the mathematical space.

[0031] First, construct an input data structure that meets the requirements of the SOMP algorithm: merge the multi-scale structural details of the simulated panchromatic component and the edge and texture details of the high-resolution image obtained in step S214 into a joint observation matrix. The multi-scale structural details of the simulated panchromatic component and the edge and texture details of the high-resolution image in the same joint observation matrix correspond to the feature vectors of the same core spatial location, so that the subsequent SOMP algorithm can process the two signals simultaneously.

[0032] The coefficients are determined through iterative computation using the SOMP algorithm: by determining the row norm of the inner product correlation matrix of all atoms in the overcomplete dictionary, the combined contribution of each dictionary atom to the two input signals is evaluated, thereby determining the index of the most energetic common atom in the dictionary. This ensures that, even when there are differences in numerical intensity between the hyperspectral image and the high-resolution image, the exact same dictionary primitive sequence must be used to describe the core structure, thus determining the sparse support set shared by both. After determining the common atom, the selected atom is added to the sparse support set matrix, and the original joint observation matrix is ​​projected onto the subspace spanned by the sparse support set using the least squares method, thereby calculating the specific projection weights, i.e., the sparse representation coefficients.

[0033] In some embodiments, in step S223, a joint sparse signal model can be used to fuse the sparse representation coefficients. By performing data fusion and image reconstruction using a joint sparse signal model, the core microstructure and mineral edge information in the high-resolution image can be transferred to the final fusion result to the greatest extent while preserving the basic spectral structure. This reduces the influence of redundant information between signals during processing, thereby preserving more spatial detail information.

[0034] Based on the sparse coefficient vectors of the corresponding hyperspectral image structural features and the corresponding high-resolution image texture features output in step S222, the L1 norm value of each non-zero element in the sparse coefficients within the neighborhood window is determined. In deep core image processing, a larger L1 norm usually corresponds to high-frequency information such as significant brightness abrupt changes, fracture edges, or mineral crystal boundaries in the image. Therefore, determining the input and discrimination criteria for the fusion process in the above manner can represent a richer amount of information.

[0035] The regional energy magnitudes of hyperspectral sparse coefficients and high-resolution sparse coefficients at the same location are compared one by one, and fused sparse representation coefficients are generated based on the absolute value maximization rule (Max-L1). Since high-resolution images are not limited by the light flux of spectral dispersion during acquisition, their texture details are usually much richer than those of hyperspectral images. Therefore, in most areas with rich core texture, this fusion method automatically identifies and selects the high-energy high-resolution sparse coefficients as the dominant ones, while retaining spectral features in areas with smooth texture. This constructs a fused coefficient vector that contains clear geometric structure and is compatible with spectral consistency, thereby logically eliminating and suppressing redundant noise in the source image, while sensitively capturing the most significant geological texture details.

[0036] In some embodiments, step S23 further includes the following steps: S231: determining the spatial detail information of each band of the hyperspectral image data based on the spatial detail information and the interpolation coefficients; S232: determining the hyperspectral high-resolution image based on the spatial detail information of all bands of the hyperspectral image data. This can improve spatial resolution, thereby increasing the accuracy of mineral identification.

[0037] In some embodiments, in step S30, the complete core image can be automatically divided into several continuous, non-overlapping strip-shaped regions at 5-10 cm intervals, based on the physical morphology of the core and the drilling direction, using a predetermined method. Each strip region serves as an independent analysis unit, and its size can be adaptively adjusted. Compared to pixel-level analysis units, segmenting and fusing image data in the above manner can avoid information overload and improve processing efficiency and stability.

[0038] In some embodiments, in step S40, the geometric coordinates of the boundaries of the segmentation results from step S30 can be extracted, and a file conforming to the Geographic Information System (GIS) standard can be generated to store the polygonal boundaries of the area and reserve attribute fields for subsequent storage of representative spectral features, mineral identification results, confidence levels, and other information of the area. Within the corresponding stripe area of ​​each file, the spectral vectors of all pixels contained in the area are statistically aggregated to generate a single, stable representative spectrum that can represent the overall spectral characteristics of the area. By generating a representative spectrum, the amount of data can be compressed, and the boundary fragmentation and computational redundancy problems caused by pixel-by-pixel recognition can be avoided.

[0039] In some embodiments, in step S40, a representative spectrum can be generated by aggregating the robust mean spectrum, median spectrum, or PCA principal component spectrum, thereby improving the stability of the representative spectrum against local outliers and noise.

[0040] In some embodiments, in step S50, the extracted regional representative spectrum can be matched and identified with a pre-set standard mineral spectrum library to improve the speed of online identification.

[0041] In some embodiments, in step S50, a combination of spectral angle matching (SAM) and spectral information divergence (SID) can be used to achieve matching identification. The two are weighted and fused (SID-SAM) to form a more stable discrimination criterion and comprehensively evaluate the differences in spectral angle and spectral distribution.

[0042] In some embodiments, in step S50, a physical rule base can be used to perform a secondary verification of the matching and identification results. This rule base contains hard threshold constraints on indicators such as the center wavelength, spectral depth, and spectral ratio of the absorption bands of specific minerals, thereby further improving the distinguishability of easily confused minerals (such as chlorite, sericite, carbonates, etc.).

[0043] Figure 3 This is a schematic diagram showing the distribution of altered minerals at different depths, obtained using the identification method provided in the embodiments of this application. Figure 3 As shown, different colored patches represent different mineral categories: yellow patches represent medium-aluminum sericite, dark blue patches represent low-aluminum sericite, light blue patches represent kaolinite, purple patches represent carbonates, and green patches represent chlorite. Because the method provided in the embodiments of this application fuses hyperspectral and high-resolution images and injects spatial details based on sparse representation into the hyperspectral image data, a high-resolution hyperspectral image is obtained. Figure 3 The mineral identification results do not present the coarse mosaic pattern commonly seen in traditional hyperspectral identification, but rather a refined distribution characteristic that closely matches the actual physical texture of the core. Figure 3 The identification results show that even in complex areas where there are fine cracks on the core surface or where minerals occur in vein-like patterns, the identification method provided in this application can accurately delineate the spatial boundaries of minerals, effectively solving the problem of misjudgment caused by spectral mixing and realizing high-precision automatic cataloging of core minerals.

[0044] Figure 4 This is a schematic diagram showing the distribution of alteration minerals in a favorable mineralization zone, obtained using the identification method provided in the embodiments of this application. Figure 4 As shown, by mapping the mineral identification results of individual rock cores to three-dimensional space or well depth profiles and calculating the alteration intensity, a geological visualization model can be intuitively constructed. Figure 4 The change from blue to red in the image indicates a gradual increase in alteration intensity; that is, the warmer the hue, the higher the alteration intensity. This model clearly reveals the enrichment patterns of hydrothermal alteration minerals in deep strata, thus assisting geologists in quickly delineating the favorable mineralization sites marked by the red boxes in the image. This demonstrates that the identification method provided in this application not only improves the accuracy of single-point identification but also provides macroscopic and quantitative geological decision-making basis for deep-earth resource exploration.

[0045] This application, in another aspect, provides a mineral classification system for rock cores, comprising: a sensor configured to acquire rock cores from deep boreholes; and a processor configured to: acquire hyperspectral image data and high-resolution image data of the rock cores; process the data to improve the spatial resolution of the hyperspectral image data to obtain fused hyperspectral high-resolution image data of the rock cores; segment the fused image data; determine a representative spectrum for each segment of the segmented image data; determine the mineral classification of each segment based on the representative spectrum; and determine the mineral classification of the rock core based on the mineral classification of each segment.

[0046] The mineral classification identification system for rock cores provided in this application improves the quality and accuracy of source data for mineral identification by obtaining high-spectral, high-resolution fused image data of the rock cores. This alleviates the problem of weakening and shifting of characteristic spectral bands caused by the mixing of multiple minerals within a pixel, enhances the distinguishability of mineral absorption bands, and thus more accurately distinguishes different minerals with similar spectral characteristics, reducing misjudgments. By segmenting the fused image data and generating representative spectra, it can replace pixel-by-pixel processing to reduce computational load and improve identification efficiency.

[0047] In some embodiments, the processor is further configured to: synchronously extract and align the spatial texture features of high-resolution image data with the spatial structural features of hyperspectral image data to obtain first data; process the first data based on sparse representation theory to obtain second data, wherein the second data is spatial detail information of joint sparse representation; and inject the spatial detail information into the hyperspectral image data to obtain a hyperspectral high-resolution image. This enables the main features of high-resolution image data and hyperspectral image data to be represented as a sparse combination, maintaining accuracy while reducing redundant information, resulting in a more concise representation.

[0048] This application also provides a computer storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, perform the aforementioned method.

[0049] Regarding the embodiments of this application, it should also be noted that, without conflict, the embodiments of this application and the features in the embodiments can be combined with each other to obtain new embodiments.

[0050] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. The scope of protection of this application shall be determined by the scope of the claims.

Claims

1. A method for identifying mineral categories in rock cores, characterized in that, It includes the following steps: S10: Obtain core samples from deep boreholes; S20: Acquire hyperspectral image data and high-resolution image data of the core, process the data to improve the spatial resolution of the hyperspectral image data, and obtain hyperspectral high-resolution fused image data of the core; S30: Segment the fused image data obtained in step S20; S40: Determine the representative spectrum of each part of the segmented image data; S50: Determine the mineral category for each part based on the representative spectrum; S60: Determine the mineral category of the core according to the mineral category of each part.

2. The identification method according to claim 1, characterized in that, Step S20 also includes the following steps: S21: Simultaneously extract and align the spatial texture features of the high-resolution image data with the spatial structural features of the hyperspectral image data to obtain the first data; S22: Process the first data based on sparse representation theory to obtain the second data, which is the spatial detail information of the joint sparse representation; S23: Inject the spatial detail information into the hyperspectral image data to obtain a hyperspectral high-resolution image.

3. The identification method according to claim 2, characterized in that, Step S21 also includes the following steps: S211: Perform IHS transformation on the hyperspectral image data to obtain the analog panchromatic components of the hyperspectral image data; S212: Perform histogram matching between the high-resolution image and the simulated panchromatic component to obtain the matched simulated panchromatic component and the high-resolution image. S213: Perform discrete wavelet transform on the simulated panchromatic components to obtain the multi-scale structural details of the simulated panchromatic components; S214: Perform edge-preserving multi-scale texture extraction processing on the high-resolution image to obtain edge and texture detail information of the high-resolution image; The first data includes the multi-scale structural details of the simulated panchromatic components of the hyperspectral image data and the edge and texture detail information of the high-resolution image.

4. The identification method according to claim 3, characterized in that, Step S22 also includes the following steps: S221: Construct a complete dictionary using the multi-scale structural details of the simulated panchromatic components in the first data and the edge and texture detail information of the high-resolution image in the first data; S222: Using the overcomplete dictionary, determine the sparse representation coefficients of the multi-scale structural details and the edge and texture details information, respectively; S223: The sparse representation coefficients are fused using the L1 norm maximization rule based on regional energy weighting to obtain the second data.

5. The identification method according to claim 4, characterized in that, In step S221, an overcomplete dictionary is constructed using the SVD algorithm.

6. The identification method according to claim 4, characterized in that, In step S222, the SOMP algorithm is used to determine the sparse representation coefficients.

7. The identification method according to claim 4, characterized in that, In step S223, the sparse representation coefficients are fused using a joint sparse signal model.

8. The identification method according to claim 2, characterized in that, Step S23 also includes the following steps: S231: Determine the spatial detail information of each band of the hyperspectral image data based on the spatial detail information and the interpolation coefficients; S232: Determine the hyperspectral high-resolution image based on the spatial detail information of all bands of the hyperspectral image data.

9. A system for identifying mineral categories in rock cores, characterized in that, It includes: A sensor configured to acquire core samples from deep boreholes; The processor is configured to: The hyperspectral image data and high-resolution image data of the core are acquired, and the data are processed to improve the spatial resolution of the hyperspectral image data, thereby obtaining hyperspectral high-resolution fused image data of the core. The fused image data is segmented; Determine the representative spectrum of each part of the segmented image data; Based on the representative spectrum, determine the mineral category for each part; The mineral category of the core is determined based on the mineral category of each part.

10. The identification system according to claim 9, characterized in that, The processor is also configured to: Simultaneously extract and align the spatial texture features of the high-resolution image data with the spatial structural features of the hyperspectral image data to obtain the first data; The first data is processed based on sparse representation theory to obtain the second data, which is the spatial detail information of the joint sparse representation. The spatial detail information is injected into the hyperspectral image data to obtain a hyperspectral high-resolution image.

11. A computer storage medium, characterized in that, The computer storage medium stores a computer program, the computer program including program instructions, which, when executed by a processor, perform the method as described in any one of claims 1-8.

Images