Mineral analysis method, device, equipment and medium

Abstract

The invention discloses a mineral analysis method, device and equipment and a medium, and relates to the technical field of fine evaluation of rock debris particle minerals. Comprising the following steps: performing morphological boundary condition identification on each rock debris particle in a segmented rock debris particle distribution diagram obtained by processing a scanned rock debris particle distribution diagram of a rock debris sample by using an Otsum threshold segmentation method to obtain a matrix pixel region; determining mineral content proportions corresponding to minerals of different mineral types corresponding to the rock debris particles in the matrix area according to the mineral distribution pixel map corresponding to the rock debris sample; determining the lithology type based on the main component mineral type of the rock debris particles, and training a decision tree based on the lithology type of the rock debris particles and the corresponding logging data; and training a small sample learning neural network based on the lithology type, the logging data and the mineral content ratio so as to determine the content ratio of each target mineral of the target rock particles based on the trained decision tree and the trained small sample learning neural network. And rock minerals can be accurately analyzed.

Description

Technical Field

[0001] The invention relates to the technical field of fine evaluation of rock cuttings particle minerals, and in particular to a mineral analysis method, device, equipment and medium. Background Art

[0002] In the fields of geoscience and petroleum exploration, in the field of well logging, the careful study of the rock skeleton is a key link in understanding and evaluating reservoir characteristics. The mineral composition of the rock skeleton not only determines the physical and chemical properties of the reservoir, such as porosity, permeability and chemical stability, but also has a direct impact on the oil content, storage capacity and ultimately the development potential of the reservoir. One of the main evaluation methods of the rock skeleton is to carry out mineral analysis. Therefore, the accurate identification and quantitative analysis of the mineral composition in the reservoir rock has become a basic and important task in oil and gas exploration and production.

[0003] Quantitative identification of rock fragments is a critical technology. Through the accurate identification and quantification of minerals in rock samples, researchers can better understand the underground geological structure, evaluate the potential value of oil and gas reservoirs, and predict the production of oil and gas. Traditional mineral identification methods rely on microscopic observation and expert experience, which is not only time-consuming and labor-intensive, but also difficult to achieve large-scale automated processing.

[0004] With the development of technology, mineral identification technology based on image processing and computer vision has begun to be studied and applied, aiming to identify and calculate the content of different minerals by automatically analyzing rock sample images. These methods usually rely on gridding the sampling area and then analyzing the mineral composition within each grid. However, this method has a significant limitation: it cannot effectively handle irregular shapes in rock samples, such as cracks, pores, and edges, which are often common in nature. Due to the existence of these irregular parts, the grid-based mineral content calculation cannot accurately reflect the actual situation, resulting in the accuracy of mineral quantitative analysis being affected.

[0005] At present, the commonly used mineral analysis method is to use an electronic probe to identify the element composition within a certain range, and then infer the mineral type based on the element data, and then obtain the mineral type of the local area. In order to uniformly test the rock cuttings, the scanned rock cuttings surface is generally divided into regions using square grids of the same size. Different grids are filled with corresponding colors according to the test results to represent different minerals, and finally the mineral type statistics are completed by the number of various colors. However, the above-mentioned mineral analysis method does not take into account the impact of rock boundary morphology and grid size changes on the calculation results. Specifically, first, it cannot be applied to the variable rock cuttings morphology. Due to the irregular external morphology of rock cuttings particles and the presence of internal pores and cracks, when a standard scale grid is used for rock cuttings segmentation, the morphological boundary represented by the boundary will be severely jagged, which cannot reflect the accurate boundary characteristics and affect the statistics of the mineral content ratio; second, the calculation method is not accurate enough. The original mineral proportion statistical method uses the number proportion method to define the mineral proportion because the grid size is the same. However, when considering the boundary effect, some grids become unstructured and their number is not enough to reflect the mineral content. Instead, area weighting should be used to calculate the mineral proportion. Third, the method of calculating the rock skeleton mineral content through logging curves and empirical formulas is not accurate enough.

[0006] In summary, how to accurately analyze rock minerals is a problem that needs to be solved urgently. Summary of the invention

[0007] In view of this, the purpose of the present invention is to provide a mineral analysis method, device, equipment and medium, which can accurately analyze rock minerals. The specific scheme is as follows:

[0008] In a first aspect, the present application discloses a mineral analysis method, comprising:

[0009] The rock cutting particle distribution map after scanning corresponding to the rock cutting sample is processed by using Otsus threshold segmentation to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map;

[0010] Performing morphological boundary condition identification on each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area, and determining the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtaining the mineral content ratios of different minerals in the particles according to the mineral pixel ratios;

[0011] Training a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles; the lithology type is determined by several main component mineral types of the rock cutting particles;

[0012] Based on the lithology type, the logging data and the mineral content ratio, a small sample learning neural network is trained to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and the content ratio of each type of mineral in the target rock particles is determined based on the trained small sample learning neural network and the target lithology type.

[0013] The method of using Otsus threshold segmentation to process the scanned rock cuttings particle distribution map corresponding to the rock cuttings sample to segment each rock cutting particle to obtain a segmented rock cuttings particle distribution map includes:

[0014] The mineral distribution pixel map corresponding to the rock cuttings sample is processed by Otsus threshold segmentation to segment each rock cutting particle, and morphological corrosion operation and morphological expansion operation are performed on each rock cutting particle to obtain the rock cutting particle distribution map after segmentation.

[0015] The step of performing morphological boundary condition recognition on each rock chip particle in the segmented rock chip particle distribution map to obtain a matrix pixel region includes:

[0016] The Canny contour detection method is used to identify the external contour boundary of each rock cutting particle in the segmented rock cutting particle distribution map, and the pixel threshold truncation method is used to identify the internal pore boundary of each rock cutting particle in the segmented rock cutting particle distribution map, so as to complete the morphological boundary condition identification of each rock cutting particle in the segmented rock cutting particle distribution map to obtain the matrix pixel area.

[0017] Wherein, before determining the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample and obtaining the mineral content ratios of different minerals in the particles according to the mineral pixel ratios, the method further includes:

[0018] The pixels corresponding to the missing mineral types at the edges or corners of the mineral distribution pixel map are supplemented based on the K-dimensional tree.

[0019] Wherein, before the scanning rock cutting particle distribution map corresponding to the rock cutting sample is processed by Otsus threshold segmentation to segment each rock cutting particle to obtain the segmented rock cutting particle distribution map, the method further includes:

[0020] The sample surface of the rock cuttings sample is treated with conductivity by using a carbon spraying technology, and the rock cuttings sample after the conductive treatment is scanned by using a scanning electron microscope to obtain the rock cuttings particle distribution map after the scan.

[0021] The step of training a small sample learning neural network based on the lithology type, the well logging data and the mineral content ratio includes:

[0022] Based on the lithology type, the logging data and the mineral content ratio, a reconstruction dimensionality reduction network module and a small sample network module in a small sample learning neural network are trained; wherein the reconstruction dimensionality reduction network module is used to perform dimensionality reduction reconstruction based on the identified lithology type to obtain a reconstructed content; and the small sample network module is used to process the reconstructed content through a fully connected network to obtain the mineral content ratio corresponding to the identified lithology type.

[0023] In a second aspect, the present application discloses a mineral analysis device, comprising:

[0024] A segmentation module, used for processing the scanned rock cutting particle distribution map corresponding to the rock cutting sample by using Otsus threshold segmentation to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map;

[0025] An identification module, used for performing morphological boundary condition identification on each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel region;

[0026] A mineral determination module is used to determine the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtain the mineral content ratios of different minerals in the particles according to the mineral pixel ratios;

[0027] A first training module is used to determine the lithology type based on the main component mineral type of the rock cutting particles, and to train a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles;

[0028] The second training module is used to train a small sample learning neural network based on the lithology type, the logging data and the mineral content ratio, so as to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and determine the content ratio of each type of mineral in the target rock particles based on the trained small sample learning neural network and the target lithology type.

[0029] Among them, the recognition module is specifically used to use the Canny contour detection method to identify the external contour boundaries of each rock cutting particle in the segmented rock cutting particle distribution map, and use the pixel threshold truncation method to identify the internal pore boundaries of each rock cutting particle in the segmented rock cutting particle distribution map, so as to complete the morphological boundary condition recognition of each rock cutting particle in the segmented rock cutting particle distribution map to obtain the matrix pixel area.

[0030] In a third aspect, the present application discloses an electronic device, including:

[0031] Memory, used to store computer programs;

[0032] A processor is used to execute the computer program to implement the mineral analysis method disclosed above.

[0033] In a fourth aspect, the present application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the mineral analysis method disclosed above is implemented.

[0034] It can be seen that the present application uses Otsus threshold segmentation to process the scanned rock cuttings particle distribution map corresponding to the rock cuttings sample to segment each rock cutting particle to obtain a segmented rock cuttings particle distribution map; morphological boundary conditions of each rock cutting particle in the segmented rock cuttings particle distribution map are identified to obtain a matrix pixel area, and the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area are determined according to the mineral distribution pixel map corresponding to the rock cuttings sample, and the mineral content ratios of different minerals in the particles are obtained according to the mineral pixel ratios; the lithology type is determined based on the main component mineral type of the rock cutting particles, and a decision tree is trained based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles; a small sample learning neural network is trained based on the lithology type, the logging data and the mineral content ratio, so as to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and determine the content ratios of various minerals in the target rock particles based on the trained small sample learning neural network and the target lithology type. It can be seen that the present application identifies the morphological boundary conditions of each rock chip particle in the rock chip particle distribution map after segmentation, reduces the influence of irregular shapes such as cracks, pores and edges, and is conducive to improving the accuracy of subsequent analysis; the present application uses decision trees and small sample learning neural networks for mineral analysis, reduces the influence of empirical reasons on the analysis results, and improves the accuracy of mineral analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0036] Figure 1 A flow chart of a mineral analysis method disclosed in this application;

[0037] Figure 2 A distribution diagram of rock cuttings particles after scanning disclosed in the present application;

[0038] Figure 3 A pixel map of mineral distribution after scanning disclosed in this application;

[0039] Figure 4 A distribution diagram of rock cuttings particles after segmentation disclosed in the present application;

[0040] Figure 5 A matrix pixel area image disclosed in the present application;

[0041] Figure 6 Another matrix pixel region image disclosed in the present application;

[0042] Figure 7 A flow chart of a specific mineral analysis method disclosed in this application;

[0043] Figure 8 A schematic diagram of a small sample learning neural network disclosed in this application.

[0044] Fig. 9 This is a schematic diagram of the structure of a mineral analysis device disclosed in this application;

[0045] Fig.10 This is a structural diagram of an electronic device disclosed in this application. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] At present, the commonly used mineral analysis method is to use an electronic probe to identify the element composition within a certain range, and then infer the mineral type based on the element data, and then obtain the mineral type of the local area. In order to uniformly test the rock cuttings, the scanned rock cuttings surface is generally divided into regions using square grids of the same size. Different grids are filled with corresponding colors according to the test results to represent different minerals, and finally the mineral type statistics are completed by the number of various colors. However, the above-mentioned mineral analysis method does not take into account the impact of rock boundary morphology and grid size changes on the calculation results. Specifically, first, it cannot be applied to the variable rock cuttings morphology. Due to the irregular external morphology of rock cuttings particles and the presence of internal pores and cracks, when a standard scale grid is used for rock cuttings segmentation, the morphological boundary represented by the boundary will be severely jagged, which cannot reflect the accurate boundary characteristics and affect the statistics of the mineral content ratio; second, the calculation method is not accurate enough. The original mineral proportion statistical method uses the number proportion method to define the mineral proportion because the grid size is the same. However, when considering the boundary effect, some grids become unstructured and their number is not enough to reflect the mineral content. Instead, area weighting should be used to calculate the mineral proportion. Third, the method of calculating the rock skeleton mineral content through logging curves and empirical formulas is not accurate enough.

[0048] To this end, the embodiment of the present application proposes a mineral analysis solution that can accurately analyze rock minerals.

[0049] The present application discloses a method for mineral analysis. Figure 1 As shown, the method includes:

[0050] Step S11: using Otsus threshold segmentation to process the scanned rock cutting particle distribution map corresponding to the rock cutting sample to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map.

[0051] In this embodiment, when analyzing the minerals, the rock cuttings samples are first screened to select representative rock cuttings that can accurately reflect the overall rock mass characteristics. Subsequently, the larger rock samples are cut into small pieces suitable for scanning electron microscope (SEM) analysis using precision rock cutting equipment. Thereafter, fine grinding and polishing steps are performed to ensure that the sample surface reaches the required smoothness and flatness for high-resolution electron microscopic scanning. In view of the inherent non-conductive nature of the rock, carbon spraying technology is used to make the sample surface conductive to avoid the accumulation of electron beams during the scanning process, which leads to a decrease in image quality. The sample preparation effect can be seen in Figure 2 and Figure 3 As shown, Figure 2 It is a distribution map of rock cuttings particles after scanning. Figure 3It is a mineral distribution pixel map after scanning. Specifically, before the rock cutting particle distribution map after scanning corresponding to the rock cutting sample is processed by Otsus threshold segmentation to segment each rock cutting particle to obtain the segmented rock cutting particle distribution map, it also includes: using carbon spraying technology to conduct conductive treatment on the sample surface of the rock cutting sample, and using a scanning electron microscope to scan the rock cutting sample after conductive treatment to obtain the rock cutting particle distribution map after scanning.

[0052] In this embodiment, the use of Otsus threshold segmentation to process the scanned rock particle distribution map corresponding to the rock cuttings sample to segment each rock cutting particle to obtain the segmented rock cutting particle distribution map includes: using Otsus threshold segmentation to process the scanned rock cutting particle distribution map corresponding to the rock cuttings sample to segment each rock cutting particle, and performing morphological corrosion operation and morphological expansion operation on each rock cutting particle to obtain the segmented rock cutting particle distribution map. It should be pointed out that after the morphological corrosion operation and the morphological expansion operation, the Canny contour detection method needs to be used to detect the edge of the segmented rock particles so that the particle edge is clear.

[0053] It should be pointed out that the Otsus threshold segmentation method (Otsus threshold segmentation) is applied to automatically determine the optimal threshold to separate rock particles from the background (i.e., white gaps). The Otsus method automatically calculates the optimal threshold by minimizing the inter-class variance. Then, for some noise in the segmentation results or small gaps between particles that are not completely separated, morphological operations erosion and dilation are used for additional processing. The erosion operation can help remove small noise points, while the dilation operation can help fill small gaps between particles so that some very small rock particles can be integrated together or integrated into large particles. Finally, the Canny contour detection method is used to detect the edges of the segmented rock particles. For details, see Figure 4 The figure shows a distribution diagram of rock cutting particles after segmentation; there are multiple rock cutting particles in the figure.

[0054] It should be pointed out that the Otsus threshold segmentation method used is a segmentation method for grayscale images.

[0055] Step S12: performing morphological boundary condition identification on each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area, and determining the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtaining the mineral content ratios of different minerals in the particles according to the mineral pixel ratios.

[0056] In this embodiment, the morphological boundary condition identification of each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area includes: using a Canny contour detection method to identify the external contour boundary of each rock cutting particle in the segmented rock cutting particle distribution map, and using a pixel threshold truncation method to identify the internal pore boundary of each rock cutting particle in the segmented rock cutting particle distribution map, so as to complete the morphological boundary condition identification of each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area.

[0057] It should be pointed out that the morphological boundary conditions of each sample are identified based on the distribution map of rock cuttings after segmentation. The boundary conditions here are divided into two parts: one is the external contour, which is identified by the Canny contour detection method, and the other is the internal pore boundary. For the external contour boundary, each rock cutting particle after segmentation is processed separately to obtain the morphological boundary of a single rock cutting particle. For the internal pore boundary, a pixel threshold-based truncation method is used to define the area where the grayscale does not meet 30 as a pore, and the mask work is used instead of the contour characterization to achieve the elimination of the non-matrix part. For details, see Figure 5 The figure shows a matrix pixel area image. The black area is the background, and the white line part is the edge of the identified non-matrix area, which is used to constrain the shape of mineral identification. Figure 6 As shown, another matrix pixel area image, different depth colors represent different minerals. It should be noted that the present application is based on the combination of Canny and threshold truncation methods to extract the effective matrix area of ​​rock cutting particles for mineral calculation.

[0058] It should be pointed out that the present application identifies and constrains the effective range of the matrix based on the boundary morphology and internal hole and crack morphology of a single particle, thereby improving the statistical accuracy of the effective area of ​​the rock cuttings particles.

[0059] In this embodiment, the method further includes: supplementing the pixels corresponding to the missing mineral types at the edges or corners of the mineral distribution pixel map based on the K-dimensional tree before determining the proportion of mineral pixels of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting samples, and obtaining the proportion of mineral content of different minerals in the particles according to the mineral pixel proportions.

[0060] It should be pointed out that, because the boundaries of rock cuttings particles are smooth during electron microscope scanning, and mineral identification is in the form of a grid, the edges often appear jagged. When smoothing the edges, there will be missing pixels in the concave part of the grid. Therefore, this application adopts a nearby reference method based on a K-dimension tree to supplement the missing local mineral type annotations. Then, taking pixels as units, first query the corresponding mineral type based on the color information of the pixel points of the mineral scanning results, as shown in Table 1, and then calculate the mineral content and proportion inside each grid based on the type and number of pixels in each grid.

[0061] Table 1 Dictionary of rock chip scanning colors and minerals

[0062] Mineral Chinese name Mineral English name Mineral Density Corresponding color code Biotite Biotite 3.01 660000 Calcite Calcite 2.78 c0c0ff Chlorite Chlorite 2.94 556b2f dolomite Dolomite 2.84 6868ff Epidote Epidote 3.42 00ff00 Iron oxide Fe-oxide 4.77 800000 Salt rock-silicate mixture Halite_Silicate_Mix 2.17 000082 Illite Illite1 2.81 00c000 Ilmenite Ilmenite 4.65 cd5c5c Kaolinite Kaolinite 2.62 804000 muscovite Muscovite 2.76 004000 Orthoclase Orthoclase 2.56 008080 Pyrite Pyrite 4.89 400000 quartz Quartz 2.63 ffc0c0

[0063] Step S13: determining the lithology type based on the main component mineral type of the rock cutting particles, and training a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles.

[0064] In a specific embodiment, the lithology type of the quartz minerals is more than 50% sandstone, and the lithology type of the dolomite and calcite minerals is more than 50% carbonate.

[0065] Step S14: Based on the lithology type, the logging data and the mineral content ratio, a small sample learning neural network is trained to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and the content ratio of each type of mineral in the target rock particles is determined based on the trained small sample learning neural network and the target lithology type.

[0066] In this embodiment, mineral classification prediction is completed based on the decision tree, and combined with the small sample learning method, the logging curve classification prediction of mineral type and content is realized.

[0067] It can be seen that the present application uses Otsus threshold segmentation to process the scanned rock cuttings particle distribution map corresponding to the rock cuttings sample to segment each rock cutting particle to obtain a segmented rock cuttings particle distribution map; morphological boundary conditions of each rock cutting particle in the segmented rock cuttings particle distribution map are identified to obtain a matrix pixel area, and the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area are determined according to the mineral distribution pixel map corresponding to the rock cutting sample, and the mineral content ratios of different minerals in the particles are obtained according to the mineral pixel ratios; a decision tree is trained based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles; the lithology type is determined by several main component mineral types of the rock cutting particles; based on the lithology type, the logging data and the mineral content ratio, a small sample learning neural network is trained to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and the content ratios of various minerals in the target rock particles are determined based on the trained small sample learning neural network and the target lithology type. It can be seen that the present application identifies the morphological boundary conditions of each rock chip particle in the rock chip particle distribution map after segmentation, reduces the influence of irregular shapes such as cracks, pores and edges, and is conducive to improving the accuracy of subsequent analysis; the present application uses decision trees and small sample learning neural networks for mineral analysis, reduces the influence of empirical reasons on the analysis results, and improves the accuracy of mineral analysis.

[0068] The embodiment of this application discloses a specific mineral analysis method. Compared with the previous embodiment, this embodiment further illustrates and optimizes the technical solution. Figure 7 As shown, specifically including:

[0069] Step S21: using Otsus threshold segmentation to process the scanned rock cutting particle distribution map corresponding to the rock cutting sample to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map.

[0070] Step S22: performing morphological boundary condition identification on each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area, and determining the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtaining the mineral content ratios of different minerals in the particles according to the mineral pixel ratios.

[0071] Step S23: determining the lithology type based on the main component mineral type of the rock cutting particles, and training a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles.

[0072] It should be pointed out that due to the strong uncertainty of using well logging to directly predict mineral types, this time a small sample learning method based on principal component mineral classification is used to accurately predict the mineral type and composition from well logging. First, the principal components of these rock cuttings are extracted, and these main minerals are used as the basis for the classification of rock cuttings. Rock cuttings with similar principal component minerals are classified into one category, and samples with different principal component minerals are placed in different types. For example, among rock cuttings, those with more than 50% quartz components are classified into the same category. Then, a classification sample library is constructed, with the principal component mineral type as the label and the well logging data of the corresponding depth section of the rock cuttings as the input. Next, the known samples are divided into training sets and test sets, and training based on decision trees is carried out. Finally, the trained decision tree is used to identify the lithology type of rock cuttings through well logging data.

[0073] Step S24: Based on the lithology type, the logging data and the mineral content ratio, the reconstruction dimension reduction network module and the small sample network module in the small sample learning neural network are trained to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and determine the content ratio of each type of mineral in the target rock particles based on the trained small sample learning neural network and the target lithology type; wherein the reconstruction dimension reduction network module is used to perform dimensionality reduction reconstruction based on the identified lithology type to obtain a reconstructed content; and the small sample network module is used to process the reconstructed content through a fully connected network to obtain the mineral content ratio corresponding to the identified lithology type.

[0074] It should be pointed out that, first, each type of rock cutting sample builds its own database. According to the depth segment corresponding to each rock cutting particle, the conventional logging curve of the depth segment plus the rock cutting lithology type obtained based on the previous step is selected as the input data, and the mineral composition and proportion are used as the output data. Each rock cutting sample corresponds to a sample. Then a small sample learning neural network is built. The network is mainly divided into two parts. The first part is the reconstruction and dimensionality reduction network module. According to the input of different types of rock cutting particle pictures, the input image is reconstructed by dimensionality reduction reconstruction to obtain the dimensionality reduction network and trained parameters. The second part is the small sample network module. First, the dimensionality reduction network and trained parameters obtained in the first step are used, and then different fully connected networks are used to process different types of rock cutting particles to generate the mineral content and proportion of this type of rock cutting particles. This method can realize cross-type knowledge transfer learning and training, extract similar features between different types of rock cutting particles, and complete independent calculations of different types of rock cutting minerals based on the unique characteristics of specific types of rock cutting particles.

[0075] See Figure 8As shown, it is a schematic diagram of a small sample learning neural network; the part copied above in the figure is the reconstructed dimensionality reduction network module, and the part reconstructed below is the small sample network module. The small sample network module also includes a fully connected neural network, in which hidden features are a more complete feature representation form, have no physical meaning, and have higher dimensions and stronger representation capabilities.

[0076] It should be pointed out that according to the small sample learning method, the mineral content prediction based on the logging data is realized based on the mineral content data statistically obtained from a small number of limited rock cuttings samples of various types.

[0077] It can be seen that the present application uses Otsus threshold segmentation to process the scanned rock cuttings particle distribution map corresponding to the rock cuttings sample to segment each rock cutting particle to obtain a segmented rock cuttings particle distribution map; morphological boundary conditions of each rock cutting particle in the segmented rock cuttings particle distribution map are identified to obtain a matrix pixel area, and the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area are determined according to the mineral distribution pixel map corresponding to the rock cuttings sample, and the mineral content ratios of different minerals in the particles are obtained according to the mineral pixel ratios; the lithology type is determined based on the main component mineral type of the rock cutting particles, and the lithology type corresponding to the rock cutting particles and the rock cutting particles are determined based on the lithology type corresponding to the rock cutting particles and the rock cutting particles. The decision tree is trained based on the logging data of the depth section; based on the lithology type, the logging data and the mineral content ratio, a small sample learning neural network is trained to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and determine the content ratio of each type of minerals in the target rock particles based on the trained small sample learning neural network and the target lithology type; wherein the reconstruction and dimensionality reduction network module is used to perform dimensionality reduction reconstruction based on the identified lithology type to obtain the reconstructed content; the small sample network module is used to process the reconstructed content through a fully connected network to obtain the mineral content ratio corresponding to the identified lithology type. It can be seen that the mineral analysis method based on the decision tree and small sample learning in this application improves the prediction accuracy of mineral content.

[0078] In a specific embodiment, in the study of a metamorphic rock well project, in order to gain a deeper understanding of the reservoir characteristics and physical properties of the key well section (3400m to 4400m), the research team used comprehensive logging technology to conduct a detailed investigation of the area. This technology includes the measurement of eight logging curves such as natural potential, acoustic wave, density, etc., and collects rock cuttings samples at a frequency of once every 2 meters for further analysis. Despite this, it has proven challenging to make a concise identification of reservoir lithology through traditional rock cutting logging methods, so electron microscope scanning and electron probe mineral composition analysis of rock cutting samples are used to promote more accurate mineral identification. The first step is to collect rock cutting samples, and complete the production of samples every 2m through a series of processes such as grinding, cementation, and carbon plating.

[0079] Then the results were analyzed. First, the particles were segmented to remove the interference of the background color. Then, each rock cutting was assigned a local area. The mineral type corresponding to the nearest pixel was used to assign unknown pixels. Then, the boundary contours and crack contours of the rock cutting particles were extracted respectively by using the method based on contour recognition and threshold truncation, and the specific effective pixel area was constrained as the effective pixel area. Finally, mineral type statistics and division were carried out based on the effective pixel data.

[0080] According to the main mineral type of each cutting particle in the cutting sample, the whole cutting is classified into 4 types, using more than 50% of the main mineral type as the classification basis. Among them, the mineral content of each type ranges from 60 to 120. These cuttings lithology types and corresponding logging curve segments are used as labels and input data, respectively, and the decision tree model is used for training to obtain the classification model.

[0081] A small sample learning sample library is constructed for the classified data, and then a small sample learning model is established. The model is trained and tested using the training and test data after the data set is divided. Finally, the target well section is predicted by first distinguishing the cuttings types based on the above decision tree model, and then taking the distinguished types as additional input data, combined with the logging data, and putting them into the small sample network model to predict the rock mineral type and proportion of the current well section.

[0082] In summary, the rock cuttings contour constraint characterization method of the present application determines a reasonable matrix range to facilitate the extraction of accurate mineral content; the pixel-based mineral proportion calculation method is more suitable for flexible boundary shapes; the mineral calculation method based on decision trees and small sample learning uses self-applicable prediction formulas for different types of minerals, thereby improving the prediction accuracy of logging curves for mineral content.

[0083] Correspondingly, the present application also discloses a mineral analysis device, see Fig. 9 As shown, the device comprises:

[0084] A segmentation module 11 is used to process the scanned rock cutting particle distribution map corresponding to the rock cutting sample by using Otsus threshold segmentation to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map;

[0085] An identification module 12 is used to identify the morphological boundary conditions of each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area;

[0086] A mineral determination module 13 is used to determine the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtain the mineral content ratios of different minerals in the particles according to the mineral pixel ratios;

[0087] A first training module 14 is used to determine the lithology type based on the main component mineral type of the rock cutting particles, and to train a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles;

[0088] The second training module 15 is used to train a small sample learning neural network based on the lithology type, the logging data and the mineral content ratio, so as to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and determine the content ratio of each type of mineral in the target rock particles based on the trained small sample learning neural network and the target lithology type.

[0089] Among them, the more specific working process of each of the above modules can refer to the corresponding content disclosed in the above embodiments, which will not be repeated here.

[0090] It can be seen that the present application uses Otsus threshold segmentation to process the mineral distribution pixel map corresponding to the rock cuttings sample to segment each rock cutting particle to obtain a segmented rock cuttings particle distribution map; morphological boundary conditions of each rock cutting particle in the segmented rock cuttings particle distribution map are identified to obtain a matrix pixel area, and the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area are determined according to the mineral distribution pixel map corresponding to the rock cuttings sample, and the mineral content ratios of different minerals in the particles are obtained according to the mineral pixel ratios; the lithology type is determined based on the main component mineral type of the rock cutting particles, and a decision tree is trained based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles; a small sample learning neural network is trained based on the lithology type, the logging data and the mineral content ratio, so as to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and determine the content ratios of various minerals in the target rock particles based on the trained small sample learning neural network and the target lithology type. It can be seen that the present application identifies the morphological boundary conditions of each rock chip particle in the rock chip particle distribution map after segmentation, reduces the influence of irregular shapes such as cracks, pores and edges, and is conducive to improving the accuracy of subsequent analysis; the present application uses decision trees and small sample learning neural networks for mineral analysis, reduces the influence of empirical reasons on the analysis results, and improves the accuracy of mineral analysis.

[0091] Furthermore, an embodiment of the present application also provides an electronic device. Fig.10 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content in the diagram cannot be regarded as any limitation on the scope of use of the present application.

[0092] Fig.10A schematic diagram of the structure of an electronic device 20 provided in an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a display screen 23, an input / output interface 24, a communication interface 25, a power supply 26, and a communication bus 27. The memory 22 is used to store a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the mineral analysis method disclosed in any of the aforementioned embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0093] In this embodiment, the power supply 26 is used to provide working voltage for each hardware device on the electronic device 20; the communication interface 25 can create a data transmission channel between the electronic device 20 and the external device, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited here; the input and output interface 24 is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0094] In addition, the memory 22, as a carrier for storing resources, may be a read-only memory, a random access memory, a disk or an optical disk, etc., and the resources stored thereon may include a computer program 221, and the storage method may be temporary storage or permanent storage. In addition to including a computer program that can be used to complete the mineral analysis method performed by the electronic device 20 disclosed in any of the aforementioned embodiments, the computer program 221 may further include a computer program that can be used to complete other specific tasks.

[0095] Furthermore, an embodiment of the present application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the mineral analysis method disclosed above is implemented.

[0096] The specific steps of the method can refer to the corresponding contents disclosed in the above embodiments, which will not be repeated here.

[0097] The various embodiments in this application are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0098] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0099] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0100] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0101] The above is a detailed introduction to a mineral analysis method, device, equipment, and storage medium provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A mineral analysis method, characterized in that: include: The rock cutting particle distribution map after scanning corresponding to the rock cutting sample is processed by using Otsus threshold segmentation to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map; Performing morphological boundary condition identification on each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel area, and determining the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtaining the mineral content ratios of different minerals in the particles according to the mineral pixel ratios; Determine the lithology type based on the main component mineral type of the rock cutting particles, and train a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles; Based on the lithology type, the logging data and the mineral content ratio, a small sample learning neural network is trained to identify the target lithology type corresponding to the target rock through the logging data based on the trained decision tree, and the content ratio of each type of mineral in the target rock particles is determined based on the trained small sample learning neural network and the target lithology type.

2. The mineral analysis method according to claim 1, characterized in that: The method of processing the scanned rock cutting particle distribution map corresponding to the rock cutting sample by using Otsus threshold segmentation to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map includes: The rock cutting particle distribution map after scanning corresponding to the rock cutting sample is processed by Otsus threshold segmentation to segment each rock cutting particle, and morphological corrosion operation and morphological expansion operation are performed on each rock cutting particle to obtain the rock cutting particle distribution map after segmentation.

3. The mineral analysis method according to claim 1, characterized in that: The step of performing morphological boundary condition recognition on each rock chip particle in the segmented rock chip particle distribution map to obtain a matrix pixel region includes: The Canny contour detection method is used to identify the external contour boundary of each rock cutting particle in the segmented rock cutting particle distribution map, and the pixel threshold truncation method is used to identify the internal pore boundary of each rock cutting particle in the segmented rock cutting particle distribution map, so as to complete the morphological boundary condition identification of each rock cutting particle in the segmented rock cutting particle distribution map to obtain the matrix pixel area.

4. The mineral analysis method according to claim 1, characterized in that: Before determining the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample and obtaining the mineral content ratios of different minerals in the particles according to the mineral pixel ratios, the method further includes: The pixels corresponding to the missing mineral types at the edges or corners of the mineral distribution pixel map are supplemented based on the K-dimensional tree.

5. The mineral analysis method according to claim 1, characterized in that: Before the scanning rock cutting particle distribution map corresponding to the rock cutting sample is processed by Otsus threshold segmentation to segment each rock cutting particle to obtain the segmented rock cutting particle distribution map, the method further includes: The sample surface of the rock cuttings sample is treated with conductivity by using a carbon spraying technology, and the rock cuttings sample after the conductive treatment is scanned by using a scanning electron microscope to obtain the rock cuttings particle distribution map after the scan.

6. The mineral analysis method according to any one of claims 1 to 5, characterized in that: The training of a small sample learning neural network based on the lithology type, the well logging data and the mineral content ratio includes: Based on the lithology type, the logging data and the mineral content ratio, a reconstruction dimensionality reduction network module and a small sample network module in a small sample learning neural network are trained; wherein the reconstruction dimensionality reduction network module is used to perform dimensionality reduction reconstruction based on the identified lithology type to obtain a reconstructed content; and the small sample network module is used to process the reconstructed content through a fully connected network to obtain the mineral content ratio corresponding to the identified lithology type.

7. A mineral analysis device, characterized in that: include: A segmentation module, used for processing the scanned rock cutting particle distribution map corresponding to the rock cutting sample by using Otsus threshold segmentation to segment each rock cutting particle to obtain a segmented rock cutting particle distribution map; An identification module, used for performing morphological boundary condition identification on each rock cutting particle in the segmented rock cutting particle distribution map to obtain a matrix pixel region; A mineral determination module is used to determine the mineral pixel ratios of different mineral types corresponding to the rock cutting particles in the matrix pixel area according to the mineral distribution pixel map corresponding to the rock cutting sample, and obtain the mineral content ratios of different minerals in the particles according to the mineral pixel ratios; A first training module is used to determine the lithology type based on the main component mineral type of the rock cutting particles, and to train a decision tree based on the lithology type corresponding to the rock cutting particles and the logging data of the depth section corresponding to the rock cutting particles; The second training module is used to train a small sample learning neural network based on the lithology type, the logging data and the mineral content ratio, so as to identify the target lithology type corresponding to the target rock particles through the logging data based on the trained decision tree, and determine the content ratio of each type of mineral in the target rock particles based on the trained small sample learning neural network and the target lithology type.

8. The mineral analysis device according to claim 7, characterized in that: The recognition module is specifically used to use the Canny contour detection method to identify the external contour boundary of each rock cutting particle in the segmented rock cutting particle distribution map, and use the pixel threshold truncation method to identify the internal pore boundary of each rock cutting particle in the segmented rock cutting particle distribution map, so as to complete the morphological boundary condition recognition of each rock cutting particle in the segmented rock cutting particle distribution map to obtain the matrix pixel area.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, used to execute the computer program to implement the mineral analysis method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: Used to store a computer program; wherein, when the computer program is executed by a processor, the mineral analysis method according to any one of claims 1 to 6 is implemented.

Images