Analysis of hyperspectral imaging data of soil particle shape and mineral composition

Abstract

The application discloses a hyperspectral imaging data processing method for analyzing rock-soil particle shape and mineral components, and comprises the following steps: acquiring hyperspectral data of a mineral particle to be measured, and preprocessing the hyperspectral data of the mineral particle to be measured; obtaining a mineral component test set based on the preprocessed hyperspectral data of the mineral particle to be measured; calling a generated mineral component model weight file to identify and classify the mineral component test set, obtaining corresponding end-member mineral components, and statistically obtaining mineral components and relative contents of each mineral component based on the end-member mineral components, so as to realize mineral component analysis of the mineral particle to be measured. The application can accurately realize joint analysis of the shape and the mineral components of the mineral particle.

Description

Technical Field

[0001] This invention relates to the field of hyperspectral data processing technology, and in particular to a hyperspectral imaging data processing method for analyzing the shape of soil and rock particles and mineral composition. Background Technology

[0002] The engineering mechanical properties of soil and rock masses are closely related to their physical and chemical characteristics, such as color, structure, chemical and mineral composition, and particle size and shape. Therefore, in fields such as engineering geological exploration, tunnel boring machine construction, and oil drilling, on-site exploration, sampling, in-situ testing, and cuttings logging are widely conducted to achieve layered identification and assessment of the engineering mechanical properties of underground soil and rock masses, guiding the design and construction of underground engineering projects. Currently, the acquisition of soil and rock samples mainly relies on manual testing methods based on visual observation and description. However, this method suffers from problems such as large sampling disturbances, strong subjectivity and reliance on experience, low efficiency, and insufficient accuracy.

[0003] In recent years, the introduction of hyperspectral imaging technology has significantly improved the ability to acquire information about soil and rock masses. For example, it has evolved from the three discrete channels of RGB (red, green, blue) color images to over 100 continuous channels of hyperspectral data. Hyperspectral imaging technology possesses continuous wide-band information acquisition capabilities in the ultraviolet, visible, and infrared bands. It utilizes spectrometers to continuously modulate the wavelength of the input light, allowing for the simultaneous acquisition of the spatial location and spectral information of the target object within the imaging area in a single imaging operation. The spectral information of the object reflects its material structure and composition characteristics, while statistical analysis of the position and distribution of pixels in two-dimensional space reveals physical attributes such as the object's color, size, and shape within the imaging area. The data acquired by hyperspectral imaging is a three-dimensional data cube, typically composed of hundreds of single-channel grayscale images superimposed. The length and width directions of the image represent two spatial dimensions, while the direction perpendicular to the image plane is called the spectral dimension.

[0004] Based on current research, hyperspectral imaging technology is mainly used for mineral composition or elemental analysis of soil and rock masses. However, the color, size, and shape characteristics of particles still rely on visual inspection or color image-based identification methods, which are technically challenging for identifying minerals with low reflectivity. To overcome the shortcomings of existing technologies and improve the level of detection, identification, and comprehensive analysis of soil and rock masses, a hyperspectral imaging data processing method for analyzing the shape of soil and rock particles and mineral composition is needed. Summary of the Invention

[0005] This invention aims to at least partially solve the technical problems in related technologies. Therefore, the object of this invention is to provide a hyperspectral imaging data processing method for analyzing the shape of rock and soil particles and mineral composition, so as to accurately achieve the joint analysis of mineral particle shape and mineral composition.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0007] A method for processing hyperspectral imaging data to analyze the shape of soil and rock particles and mineral composition includes:

[0008] Acquire hyperspectral data of the mineral particles to be tested, and preprocess the hyperspectral data of the mineral particles to be tested;

[0009] A mineral composition test set was obtained based on the preprocessed hyperspectral data of the mineral particles to be tested.

[0010] The generated mineral component model weight file is called to identify and classify the mineral component test set, obtain the corresponding end-member mineral components, and obtain the mineral components and the relative content of each mineral component based on the end-member mineral components, thereby realizing the mineral component analysis of the mineral particles to be tested.

[0011] Preferably, the hyperspectral data of the mineral particles to be tested are preprocessed, including:

[0012] The basic information of the hyperspectral data of the mineral particles to be tested is read, and the hyperspectral data of the mineral particles to be tested is categorized and corrected. The basic information includes data format, spectral range, image depth, image size and sampling interval. The data categorization and correction include region of interest delineation, baseline correction, reflectance calibration, spectral noise reduction, reflectance parameter normalization and derivative spectrum calculation.

[0013] Preferably, the preprocessed hyperspectral data of the mineral particles to be tested are decomposed into spectral vectors according to different endmembers to obtain the mineral component test set.

[0014] Preferably, the steps for generating the mineral component model weight file include:

[0015] Obtain hyperspectral data of standard mineral particles, wherein the standard mineral particles are mineral particles with known mineral composition;

[0016] A labeled array is generated and superimposed on the hyperspectral data of standard mineral particles to form labeled standard hyperspectral data;

[0017] The labeled standard hyperspectral data were decomposed into standard spectral vectors according to different endmembers, and then distributed into a mineral component training set and a mineral component validation set according to the data allocation ratio.

[0018] The mineral component training set and mineral component validation set are input into a one-dimensional deep learning model for iterative training and validation to generate the mineral component model weight file.

[0019] Preferably, the annotation array includes boundary elements and type elements;

[0020] The gray value corresponding to the boundary element is 0, it is located at the particle boundary of the standard mineral particle, and the boundary element of a single mineral particle forms a closed loop.

[0021] The gray value corresponding to the type element is in the range of 1 to 255, located inside the standard mineral particle, and the type element of a single mineral particle forms a connected domain;

[0022] The grayscale value of the type element corresponds one-to-one with the mineral composition of the standard mineral particle.

[0023] Preferably, the method further includes particle segmentation and shape analysis of the mineral particles to be tested, which includes:

[0024] A particle segmentation test set was obtained based on the preprocessed hyperspectral data of the mineral particles to be tested.

[0025] The generated particle segmentation model weight file is used to perform instance segmentation on the particle segmentation test set to obtain the instance boundary point coordinates of the mineral particles to be tested, and a two-dimensional instance mask is generated.

[0026] Connectivity analysis is performed on the two-dimensional instance mask to obtain its shape parameters, and the shape and size distribution of the individual mineral particles to be tested are statistically obtained to achieve particle segmentation and shape analysis.

[0027] Preferably, the preprocessed hyperspectral data of the mineral particles to be tested is decomposed into spectral images of the particles to be tested according to different wavelengths to obtain the particle segmentation test set.

[0028] Preferably, the steps for generating the particle segmentation model weight file include:

[0029] Synthesize feature grayscale images;

[0030] The feature grayscale image is superimposed with the annotation array, and after performing random data augmentation, it is randomly cropped into several images and annotation files of preset sizes, and then distributed according to the data allocation ratio to obtain the particle segmentation training set and particle segmentation validation set.

[0031] The particle segmentation training set and particle segmentation validation set are input into the two-dimensional deep learning model for iterative training and validation to generate the particle segmentation model weight file.

[0032] Preferably, the synthesized feature grayscale image includes:

[0033] Based on the comparative analysis of mineral component identification results, the mineral component dataset is used to determine the characteristic absorption peaks of mineral components, extract the center wavelength and full width at half maximum (FWHM) of the characteristic absorption peaks, and calculate the difference characteristic wavelengths. The mineral component dataset includes the mineral component training set and the mineral component validation set.

[0034] A differential feature wavelength search function is defined. The differential feature wavelength search is performed within the spectral range using the differential feature wavelength search function to extract the grayscale image at the center wavelength and the grayscale image at the differential feature wavelength. The extracted image is then processed using linear superposition and contrast enhancement methods to synthesize the feature grayscale image.

[0035] Preferably, the random data augmentation operation includes at least noise injection, blurring, rotation, flipping, and color transformation operations.

[0036] This invention has at least the following technical effects:

[0037] This invention provides a hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition. This method can simultaneously analyze the particle shape and mineral composition of soil particles through a single hyperspectral imaging, forming a multi-parameter data acquisition and analysis capability. It further simplifies the system structure and avoids the adverse effects of secondary measurements on the measurement and analysis speed and accuracy. At the same time, it obtains and filters particle segmentation data based on mineral composition identification and classification results, effectively improving the quantity and quality of deep learning training and validation data, and significantly reducing the annotation workload.

[0038] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0039] Figure 1 This is a flowchart of a hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition according to an embodiment of the present invention.

[0040] Figure 2 This is a flowchart illustrating the specific implementation of the hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition according to an embodiment of the present invention.

[0041] Figure 3 This is a schematic diagram of a standard sample base according to an embodiment of the present invention. Detailed Implementation

[0042] The following describes this embodiment in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the invention, and should not be construed as limiting the invention.

[0043] The following describes, with reference to the accompanying drawings, a hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition according to this embodiment.

[0044] Figure 1 This is a flowchart illustrating a hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0045] Step S101: Obtain the hyperspectral data of the mineral particles to be tested, and preprocess the hyperspectral data of the mineral particles to be tested.

[0046] The preprocessing of the hyperspectral data of the mineral particles to be tested includes reading the basic information from the hyperspectral data of the mineral particles to be tested and performing data partitioning correction on the hyperspectral data of the mineral particles to be tested. The basic information includes data format, spectral range, image depth, image size and sampling interval; the data partitioning correction includes region of interest delineation, baseline correction, reflectance calibration, spectral noise reduction, reflectance parameter normalization processing and calculation of derivative spectrum.

[0047] Step S102: Obtain a mineral component test set based on the preprocessed hyperspectral data of the mineral particles to be tested.

[0048] Specifically, the preprocessed hyperspectral data of the mineral particles to be tested are decomposed into spectral vectors according to different endmembers to obtain the mineral component test set.

[0049] Step S103: Call the generated mineral component model weight file to identify and classify the mineral component test set, obtain the corresponding end-member mineral components, and obtain the mineral components and the relative content of each mineral component based on the end-member mineral components, so as to realize the mineral component analysis of the mineral particles to be tested.

[0050] The steps for generating the mineral component model weight file include: acquiring hyperspectral data of standard mineral particles (mineral particles with known mineral composition); generating a labeled array and overlaying it onto the hyperspectral data of the standard mineral particles to form labeled standard hyperspectral data; dispersing the labeled standard hyperspectral data into standard spectral vectors according to different endmembers, and allocating them according to the data allocation ratio to obtain a mineral component training set and a mineral component validation set; inputting the mineral component training set and the mineral component validation set into a one-dimensional deep learning model for iterative training and validation to generate the mineral component model weight file.

[0051] In this embodiment, the annotation array includes boundary elements and type elements. Boundary elements have a grayscale value of 0, are located at the particle boundaries of the standard mineral particles, and the boundary elements of a single mineral particle form a closed loop. Type elements have a grayscale value in the range of 1 to 255, are located inside the standard mineral particles, and the type elements of a single mineral particle form a connected component. It should be noted that the grayscale value of the type element corresponds one-to-one with the mineral composition of the standard mineral particle.

[0052] In one embodiment of the present invention, the method further includes particle segmentation and shape analysis of the mineral particles to be tested. Particle segmentation and shape analysis of the mineral particles to be tested includes: obtaining a particle segmentation test set based on the preprocessed hyperspectral data of the mineral particles to be tested; calling the generated particle segmentation model weight file to perform instance segmentation on the particle segmentation test set to obtain the coordinates of the instance boundary points of the mineral particles to be tested, and generating a two-dimensional instance mask; performing connected component analysis on the two-dimensional instance mask to obtain the shape parameters of the two-dimensional instance mask, and statistically obtaining the single particle shape and particle size distribution information of the mineral particles to be tested, thereby realizing particle segmentation and shape analysis.

[0053] The particle segmentation test set can be obtained by dispersing the preprocessed hyperspectral data of the mineral particles to be tested into spectral images according to different wavelengths.

[0054] In one embodiment of the present invention, the step of generating the particle segmentation model weight file includes: synthesizing a feature grayscale image, superimposing the feature grayscale image with a label array, and after performing random data augmentation operations such as noise injection, blurring, rotation, flipping, and color transformation, randomly cropping it into several images and label files of preset sizes, and allocating them according to the data allocation ratio to obtain a particle segmentation training set and a particle segmentation validation set. The particle segmentation training set and the particle segmentation validation set are then input into a two-dimensional deep learning model for iterative training and validation to generate the particle segmentation model weight file.

[0055] In this embodiment, the synthesis of feature grayscale images includes: comparing and analyzing the mineral component dataset based on the mineral component identification results to determine the feature absorption peaks of the mineral components, extracting the center wavelength and full width at half maximum (FWHM) of the feature absorption peaks, so as to calculate the difference feature wavelengths. The mineral component dataset includes a mineral component training set and a mineral component validation set.

[0056] Furthermore, a differential feature wavelength search function is defined. By using the differential feature wavelength search function to search for differential feature wavelengths within the spectral range, grayscale images at the center wavelength and grayscale images at the differential feature wavelengths are extracted. Then, the extracted images are processed by linear superposition and contrast enhancement methods to synthesize the feature grayscale images.

[0057] Figure 2 This is a flowchart illustrating the specific implementation of the hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition according to an embodiment of the present invention. Figure 2 As shown, the method specifically includes:

[0058] Step 1: Hyperspectral Data Preprocessing

[0059] (1) Read basic information

[0060] The basic information in the hyperspectral data of the mineral particles to be tested includes data format, spectral range (start wavelength SW, end wavelength EW), image depth, image size, and sampling interval SI.

[0061] (2) Data partition correction

[0062] The hyperspectral data of the mineral particles to be tested are subjected to data partitioning correction, which includes delineation of the region of interest (ROI), baseline correction, reflectance calibration, spectral noise reduction, normalization, and calculation of derivative spectra.

[0063] Region of interest delineation involves removing data within the hyperspectral scanning imaging region that is irrelevant to or weakly correlated with the test sample, thereby reducing computational load and improving processing and analysis efficiency.

[0064] Baseline correction removes the influence of factors such as lenses, beam splitters, and optical fibers on the sample spectrum during light propagation, in order to reflect the true spectral information of the test sample.

[0065] Reflectivity calibration is performed using the spectral data of a black and a white standard diffuse reflectance target plate, converting the original light intensity parameters into reflectivity parameters. The black and white standard diffuse reflectance target plates are... Figure 3Reference numerals 22 and 23 are provided in the square groove of the standard sample base 21. Standard mineral particles 24 are also provided in the square groove of the standard sample base 21. A positioning target 20 is also provided on the standard sample base 21.

[0066] Spectral noise reduction uses digital signal filtering technology to reduce random noise caused by factors such as the system itself and the test environment in spectral data, thereby reducing invalid information in the spectral data.

[0067] Normalization is a process that normalizes the reflectance parameter to the [0, 1] interval to enhance the comparability of spectral data.

[0068] Calculating derivative spectra involves calculating first- and second-order derivative spectra to highlight subtle features of the spectral curves, separate similar / overlapping peaks, and eliminate baseline drift.

[0069] Step 2: Generate a deep learning dataset

[0070] (1) Data labeling

[0071] Data annotation involves generating an array of annotations and overlaying it onto the hyperspectral data of standard mineral particles (mineral particles with known mineral composition) to form an annotated standard hyperspectral data.

[0072] Furthermore, the annotation array includes boundary elements and type elements, and the size of the annotation array is the same as the image size, which is 2560px*1920px.

[0073] Furthermore, the gray value corresponding to the boundary element is 0, located at the particle boundary position of the standard mineral particle, and for a single mineral particle, its boundary element forms a closed loop.

[0074] Furthermore, the grayscale value corresponding to the type element is in the range of [1, 255], located inside the standard mineral particle, and for a single mineral particle, its type element forms a connected domain.

[0075] Furthermore, the grayscale value of the type element corresponds one-to-one with the mineral composition of the standard mineral particle.

[0076] (2) Generate mineral component dataset

[0077] The mineral composition dataset includes a mineral composition training set, a mineral composition validation set, and a mineral composition test set.

[0078] The mineral component training set and mineral component validation set are obtained by dispersing the labeled standard hyperspectral data into standard spectral vectors according to different endmembers and distributing them according to the data allocation ratio.

[0079] The mineral component test set is obtained by dispersing the preprocessed hyperspectral data of the mineral particles to be tested into spectral vectors according to different endmembers.

[0080] (3) Generate particle segmentation test set

[0081] The particle segmentation test set is obtained by dispersing the preprocessed hyperspectral data of the mineral particles to be tested into spectral images according to different wavelengths.

[0082] Step 3: Mineral Component Identification and Classification

[0083] (1) Input the mineral component training set and mineral component validation set into the one-dimensional deep learning model for iterative training and validation to obtain the mineral component model weight file.

[0084] (2) By calling the mineral component model weight file, the mineral component test set is identified and classified to obtain the corresponding end-member mineral components, and further statistical information such as mineral components and the relative content of each component is obtained.

[0085] Step 4: Particle segmentation and shape analysis

[0086] (1) Calculate the difference characteristic wavelength

[0087] Based on the comparative analysis of the mineral component dataset using the mineral component identification results, the characteristic absorption peaks of the mineral components are obtained. The center wavelength (CW) and full width at half maximum (FWHM) of the characteristic absorption peaks are extracted, and the differential characteristic wavelengths (CWL, CWR) are calculated.

[0088]

[0089] Wherein, CWL and CWR represent the left difference characteristic wavelength and the right difference characteristic wavelength, respectively.

[0090] (2) Synthesizing Feature Gray-Scale Images

[0091] Define the differential feature wavelength search function SF(x):

[0092]

[0093] Where n = INT[(x-SW)] / SI], x represents the independent variable of the differential characteristic wavelength search function SF(x), SW represents the starting wavelength of the spectral range, SI is the sampling interval, n represents the intermediate variable, which is a non-negative integer, and INT represents the floor function.

[0094] Within the spectral range, a differential characteristic wavelength search is performed to extract the grayscale image Img(SF(CW)) at the center wavelength and the grayscale images Img(SF(CWL)) and Img(SF(CWR)) at the differential characteristic wavelengths. A characteristic grayscale image SCI is then synthesized using linear superposition and contrast enhancement methods.

[0095]

[0096] Where CGI represents the fused image and HE is the histogram equalization function.

[0097] (3) Generate particle segmentation training set and particle segmentation validation set

[0098] The particle segmentation training set and particle segmentation validation set are obtained by overlaying feature grayscale images with annotation arrays, performing random data augmentation, randomly cropping them into several 224px*224px images and annotation files, and distributing them according to the data allocation ratio.

[0099] Preferably, the random data augmentation operations include noise injection, blurring, rotation, flipping, and color transformation.

[0100] (4) Particle segmentation training and verification

[0101] The particle segmentation training set and particle segmentation validation set are input into the two-dimensional deep learning model for iterative training and validation to obtain the particle segmentation model weight file.

[0102] (5) Particle segmentation test

[0103] By calling the particle segmentation model weight file, the particle segmentation test set is segmented into instances to obtain the coordinates of the instance boundary points of the mineral particles to be tested, and a two-dimensional instance mask is generated.

[0104] (6) Shape Analysis

[0105] Connectivity analysis is performed on the two-dimensional instance mask to obtain its shape parameters. Then, information such as the single particle shape and particle size distribution of the mineral particles to be tested are obtained through further statistical analysis.

[0106] In summary, this invention provides a hyperspectral imaging data processing method for analyzing the shape of soil particles and mineral composition. This method can simultaneously analyze the particle shape and mineral composition of soil particles through a single hyperspectral imaging, forming a multi-parameter data acquisition and analysis capability. This further simplifies the system structure and avoids the adverse effects of secondary measurements on the measurement and analysis speed and accuracy. At the same time, by acquiring and screening particle segmentation data based on mineral composition identification and classification results, it effectively improves the quantity and quality of deep learning training and validation data, and can significantly reduce the annotation workload.

[0107] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0108] Although the present invention has been described in detail through the preferred embodiments above, it should be understood that the above description should not be considered as a limitation of the present invention. Various modifications and substitutions to the present invention will be apparent to those skilled in the art after reading the above description. Therefore, the scope of protection of the present invention should be defined by the appended claims.

Claims

1. A hyperspectral imaging data processing method for analyzing rock-soil particle shape and mineral composition, characterized in that, The method comprises the following steps: acquiring hyperspectral data of the mineral particles to be tested, and preprocessing the hyperspectral data of the mineral particles to be tested; obtaining a mineral component test set based on the preprocessed hyperspectral data of the mineral particles to be tested; calling the generated mineral component model weight file to identify and classify the mineral component test set, obtaining the corresponding end-member mineral component, and statistically obtaining the mineral component and the relative content of each mineral component based on the end-member mineral component, thereby realizing mineral component analysis of the mineral particles to be tested; the generation of the mineral component model weight file comprises the following steps: acquiring hyperspectral data of standard mineral particles, wherein the standard mineral particles are mineral particles with known mineral components; generating a label array and superimposing it on the hyperspectral data of the standard mineral particles to form labeled standard hyperspectral data; dissolving the labeled standard hyperspectral data into standard spectral vectors according to different end-members, and distributing the standard spectral vectors according to a data distribution ratio to obtain a mineral component training set and a mineral component verification set; inputting the mineral component training set and the mineral component verification set into a one-dimensional deep learning model for iterative training and verification, so as to generate the mineral component model weight file; The method further comprises particle segmentation and shape analysis of the mineral particles to be tested, which comprises the following steps: obtaining a particle segmentation test set based on the preprocessed hyperspectral data of the mineral particles to be tested; calling the generated particle segmentation model weight file to perform instance segmentation on the particle segmentation test set, obtaining the instance boundary point coordinates of the mineral particles to be tested, and generating a two-dimensional instance mask; performing connected component analysis on the two-dimensional instance mask to obtain the shape parameters of the two-dimensional instance mask, and statistically obtaining the single-particle shape and particle size distribution information of the mineral particles to be tested, thereby realizing particle segmentation and shape analysis; the generation of the particle segmentation model weight file comprises the following steps: synthesizing a feature grayscale image; superimposing the feature grayscale image and the label array, and after random data augmentation, randomly cropping the feature grayscale image and the label array into a plurality of images and label files with a preset size, and distributing the images and the label files according to a data distribution ratio to obtain a particle segmentation training set and a particle segmentation verification set; inputting the particle segmentation training set and the particle segmentation verification set into a two-dimensional deep learning model for iterative training and verification, so as to generate the particle segmentation model weight file; synthesizing a feature grayscale image comprises the following steps: comparing and analyzing a mineral component data set based on the mineral component identification result, so as to determine the characteristic absorption peak of the mineral component, extract the center wavelength and full width at half maximum of the characteristic absorption peak, and calculate the difference characteristic wavelength, wherein the mineral component data set comprises the mineral component training set and the mineral component verification set; defining a difference characteristic wavelength search function, searching for the difference characteristic wavelength in the spectral range through the difference characteristic wavelength search function, extracting the grayscale image at the center wavelength and the grayscale image at the difference characteristic wavelength, and processing the extracted images through a linear superposition and contrast enhancement method, so as to synthesize the feature grayscale image.

2. The method of claim 1, wherein the method is characterized by: The preprocessing of the hyperspectral data of the mineral particles to be tested comprises the following steps: Basic information in hyperspectral data of the mineral particle to be tested is read, and the hyperspectral data of the mineral particle to be tested is subjected to data partition correction; wherein the basic information includes data format, spectral range, image depth, image size and sampling interval; the data partition correction includes region of interest delineation, baseline correction, reflectivity calibration, spectral noise reduction, reflectivity parameter normalization processing and derivative spectrum calculation.

3. The method of claim 1, wherein the method further comprises: The hyperspectral data of the preprocessed mineral particle to be tested is dispersed into mineral component test sets according to different end members.

4. The method of claim 1, wherein the method is used for analyzing the hyperspectral imaging data of rock and soil particles for shape and mineral composition. The label array includes boundary elements and type elements; The boundary elements correspond to a gray value of 0, are located at the particle boundary positions of the standard mineral particles, and the boundary elements of a single mineral particle form a closed loop; The type elements correspond to a gray value in a range of 1-255, are located at the interior of the standard mineral particles, and the type elements of a single mineral particle form a connected domain; The gray value of the type element corresponds to the mineral composition of the standard mineral particle in a one-to-one manner.

5. The method of claim 1, wherein the method further comprises: The hyperspectral data of the preprocessed mineral particle to be tested is dispersed into particle segmentation test sets according to different wavelengths.

6. The method of claim 1, wherein the method further comprises: The random data enhancement operation at least includes noise injection, blur processing, rotation, flipping and color transformation operations.

Images