A multi-component segmentation method for shale scanning electron microscope images

Abstract

The present invention discloses a multi-component segmentation method for shale SEM images. Based on the grayscale differences among pores, organic matter, and matrix minerals in shale SEM images, a multi-threshold segmentation method is used to perform initial multi-component calibration of the image. Based on the developmental characteristics of the organic matter, an area screening method and interactive image processing are then used to refine the organic matter. A boundary expansion algorithm is applied to the organic matter to ensure that it coincides with the organic pores, thereby screening out the organic pores. Subtraction operations are then used to calibrate the inorganic pores. Finally, Boolean operations are used to segment the shale matrix minerals, completing the multi-component segmentation of the shale SEM image. This method is simple to operate and produces highly accurate segmentation results. In addition to qualitatively segmenting organic and inorganic pores, it can also be used for quantitative calculation of different pore parameters, possessing broad application value in shale gas reservoir evaluation and exploration and development.

Description

Technical Field

[0001] The present invention relates to the technical field of nano-CT scanning and digital core image processing, and in particular to a multi-component segmentation method for shale scanning electron microscope images. Background Art

[0002] Characterizing and characterizing the micro- and nano-pore structures of shale gas reservoirs is fundamental to shale gas exploration and development, and is crucial for reservoir evaluation and fluid transport mechanisms. During shale gas exploration and development, coring and experimental methods are used to characterize the pore structure. However, due to the varying scales of the various components, quantitative evaluation and classification of complex components such as organic pores, inorganic pores, and organic matter are difficult. With the development of nanoscale scanning electron microscopy (SEM) imaging technology, intuitive characterization of the multi-component nanostructure of shale using high-resolution SEM imaging has become an important research tool.

[0003] Shale reservoirs have complex pore and fracture structures, with significant morphological differences between components and strong reservoir heterogeneity. The presence of pore and fracture backgrounds and oblique shadows in scanned grayscale images significantly affects the delineation of organic matter, making it difficult for traditional image segmentation methods to achieve refined segmentation of scanned images. Furthermore, organic and inorganic pores have similar grayscale values ​​in scanned images, making them difficult to distinguish effectively using conventional methods. Therefore, it is necessary to develop appropriate screening methods that incorporate the unique relationships between shale components to qualitatively segment organic pores, inorganic pores, and microfractures, improve the accuracy of the delineation results, and provide technical support for the quantitative characterization of pore and fracture morphology. Summary of the Invention

[0004] The present invention mainly overcomes the deficiencies in the prior art and aims to provide a multi-component classification method for shale scanning electron microscope images.

[0005] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0006] A multi-component classification method for shale scanning electron microscope images, characterized in that the calculation method comprises the following steps:

[0007] S1: Scan the polished surface of the shale sample using a nano-scanning electron microscope to obtain a two-dimensional grayscale image of the core;

[0008] S2: Preprocessing the two-dimensional core grayscale image obtained in step S1 to improve image contrast and eliminate image artifacts and noise;

[0009] S3: using a threshold segmentation method to perform a preliminary segmentation of the organic matter in the image pre-processed in step S2;

[0010] S4: superimposing the pre-processed image in step S2 and the organic matter image initially divided in step S3, changing the transparency of either image to detect over-divided areas, and performing fine correction on the over-divided areas through image editing to obtain a secondary division image of the organic matter;

[0011] S5: Combine the threshold segmentation method and the brightness difference segmentation method to segment all pores and cracks with low grayscale values, and obtain an image containing all pores and cracks. Combine the connectivity test and the aspect ratio screening criteria to filter the cracks and obtain an image containing only all cracks.

[0012] S6: removing all cracks divided in step S5 from the image of all pores and cracks in step S5 by a subtraction operation, so that the image after removal contains both organic and inorganic pores, and performing a connectivity test and numbering on all pores;

[0013] S7: Using a boundary expansion algorithm, the outer boundaries of all pores in step S6 are expanded outward by 1 pixel. The organic matter refined and corrected in step S4 and the pores after boundary expansion are superimposed on the same base map. The boundaries of the organic pores in the overlapping parts are constructed by taking the intersection method. The pores in the overlapping parts are screened to obtain all the numbers corresponding to the organic pores.

[0014] S8: Using the numbered image in step S6 as a sample space, only the pores corresponding to the organic pore numbers in step S7 are screened to obtain an image containing only organic pores; and organic pores are subtracted from the sample space by a subtraction operation to obtain an image containing only inorganic pores.

[0015] S9: taking a union of the organic matter image refined and corrected in step S4 and the image containing all pores and fractures in step S5, and taking a complement of the unioned images to obtain a shale matrix mineral image;

[0016] S10: In the same image, the organic matter after refinement and correction in step S4, the cracks screened out in step S5, the organic pores and inorganic pores screened out in step S8, and the shale matrix minerals in step S9 are superimposed respectively, and the above five components are marked with image numbers 0, 1, 2, 3, and 4, respectively, thereby achieving multi-component refinement and qualitative segmentation of the scanned grayscale image.

[0017] Furthermore, the preparation process of the polished surface of the shale sample in step S1 is as follows: first, the surface of the shale sample is rock-ground using ultra-thin emery paper until the surface is flat; then, the sample surface is polished using argon ion polishing equipment under vacuum conditions to obtain a smooth polished surface.

[0018] Furthermore, the method for improving image contrast in step S2 is contrast enhancement. Shale images scanned by a scanning electron microscope often have dark grayscale values ​​and poor image contrast. The narrow grayscale gradient makes subsequent image segmentation difficult. Contrast enhancement algorithms can proportionally amplify grayscale gradients without changing the original image information.

[0019] The noise elimination method in step S2 is non-local mean filtering. SEM images generally require filtering due to electronic signal interference. For multi-component segmentation of shale images, determining the boundaries of each mineral component in the image is particularly important. Using non-local mean filtering to filter the image can eliminate noise while preserving image boundary information as much as possible.

[0020] Furthermore, the organic matter in step S3 is mainly carbonaceous components, which often show light gray features with medium grayscale in the scanning electron microscope image. By selecting the lower and upper grayscale thresholds of the organic matter, the positioning mark of the organic matter components can be achieved.

[0021] Furthermore, in step S4, to check the accuracy of the segmentation results, the organic matter image initially segmented in step S3 and the preprocessed grayscale image from step S2 are superimposed. The transparency of the organic matter image is repeatedly varied to dynamically observe and identify inaccurately segmented areas. Excessively segmented organic matter areas primarily include shadows on polished surface bevels and the backgrounds of pores and seams. Using an image editor, the pixel values ​​of areas incorrectly labeled as organic matter are changed from 1 to 0, completing the detailed correction of the organic matter.

[0022] Furthermore, in step S5, a threshold segmentation method is used to divide pores with larger apertures and cracks with larger openings, and a brightness difference segmentation method is used to divide pores with smaller apertures and cracks with smaller openings. Finally, the results of the two segmentation methods are combined to achieve a qualitative division of all pores and cracks.

[0023] Furthermore, the connectivity test method in step S5 is specifically as follows: the same pore or crack is marked with the same number, and the marking numbers of the pores or cracks that are not connected to each other are different.

[0024] Currently, the method for classifying cracks and pores is mainly based on morphological screening methods. By calculating the aspect ratio of the minimum circumscribed rectangle of each pore and crack, combined with the threshold screening standard of the aspect ratio, pores and cracks are divided. The lower limit of the crack aspect ratio determined by this method is 4, that is, a crack is defined as a crack if the aspect ratio is greater than 4, and a pore is defined as a pore if the aspect ratio is less than 4.

[0025] Furthermore, the connectivity evaluation criteria in step S6 are as follows: for a single pixel in a two-dimensional plane, it has four adjacent pixels and four diagonal pixels. If the adjacent pixel values ​​or the diagonal pixel values ​​are consistent with the pixel value of the pixel, the two pixels are considered to be connected. Interconnected pores are marked with the same number, and unconnected pores are marked with different numbers.

[0026] Furthermore, step S7 utilizes the unique relationship between organic pores and organic matter to screen for organic pores. Since organic pores primarily arise from the thermal evolution of organic matter, the space containing them is always within the organic matter. Therefore, the outer boundaries of the organic pores and the inner boundaries of the organic matter touch each other in the image. Expanding the outer boundaries of the organic pores will overlap with the organic matter. By screening the pore numbers of the overlapping regions, the numbers of all organic pores can be obtained. However, the outer boundaries of inorganic pores generally have no correlation with the organic matter.

[0027] Furthermore, the classification of shale matrix minerals in step S9 primarily utilizes Boolean operations. Since the organic matter and all pores and fractures have already been divided and corrected in steps S4 and S5, the use of Boolean operations allows for rapid classification of shale matrix minerals, avoiding classification errors caused by shadows on the pore and fracture backgrounds and slopes.

[0028] This paper provides a multi-component classification method for shale SEM images. This method uses the morphology and distribution characteristics of different pore types as the primary classification basis, employing digital core technology and Boolean operations as the primary mathematical methods. It employs specific classification methods for different components, theoretically demonstrating a method for classifying shale SEM grayscale images into five component images. This method is simple to operate and produces highly accurate classification results. In addition to qualitatively segmenting organic and inorganic pores, it can also be used for quantitative calculation of various pore parameters, demonstrating its broad application in shale gas reservoir evaluation and exploration and development.

[0029] Beneficial effects:

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] Based on the morphology and distribution characteristics of different pore types, and using digital core technology and Boolean operations as the primary mathematical methods, this paper theoretically describes a method for classifying shale scanning electron microscope grayscale images into five component images. This method is simple to operate and produces highly accurate results. In addition to qualitatively segmenting organic and inorganic pores, it can also be used for quantitative calculation of various pore parameters, demonstrating its broad application in shale gas reservoir evaluation, exploration, and development. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a grayscale scanning electron microscope image of shale;

[0033] Figure 2 This is the primary and secondary division diagram of organic matter;

[0034] Figure 3 This is the result diagram of the division of organic pores and inorganic pores;

[0035] Figure 4 This is a comparison chart of multi-component partitioning results;

[0036] Figure 5 Schematic diagram of the multi-component classification process of shale SEM images. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail with reference to the following examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0038] Example:

[0039] A multi-component classification method for shale scanning electron microscope images includes the following steps:

[0040] S1: A two-dimensional grayscale image of the core is obtained by scanning the polished surface of the shale sample using a nanoscale scanning electron microscope. The specific process is as follows: first, the surface of the shale sample is rock-ground using ultra-thin emery paper until the surface is flat; then, the sample surface is polished using an argon ion polishing device under vacuum conditions to obtain a smooth polished surface.

[0041] S2: Preprocess the two-dimensional core grayscale image obtained in step S1 to improve image contrast and eliminate image artifacts and noise:

[0042] Contrast enhancement is a method for improving image contrast. Shale images obtained through scanning electron microscopy often have dark grayscale values, poor image contrast, and narrow grayscale gradients, which can make subsequent image segmentation difficult. Contrast enhancement algorithms can proportionally amplify grayscale gradients without changing the original image information.

[0043] The method used to eliminate noise is non-local mean filtering. SEM images generally require filtering due to electronic signal interference. For multi-component segmentation of shale images, determining the boundaries of each mineral component is particularly important. Using non-local mean filtering can eliminate noise while preserving image boundary information as much as possible.

[0044] S3: Use the threshold segmentation method to perform initial segmentation on the organic matter in the image preprocessed in step S2: Organic matter is mainly composed of carbonaceous components, which often appear as light gray features with medium grayscale in the scanning electron microscope image. By selecting the lower and upper limits of the grayscale threshold of the organic matter, the positioning mark of the organic matter components can be achieved.

[0045] S4: Superimpose the preprocessed image from step S2 with the organic matter image initially divided from step S3. Change the transparency of either image to identify over-divided areas. Fine-tune these over-divided areas through image editing to obtain a secondary organic matter image. To verify the accuracy of the division results, superimpose the organic matter image initially divided from step S3 with the preprocessed grayscale image from step S2. Dynamically observe and confirm inaccurately divided areas by repeatedly changing the transparency of the organic matter image. Excessively divided organic matter areas primarily include polished surface bevel shadows and pore backgrounds. Fine-tune these areas by changing the pixel values ​​of areas incorrectly marked as organic matter from 1 to 0 using an image editor.

[0046] S5: Combine the threshold segmentation method and the brightness difference segmentation method to segment all pores and cracks with low grayscale values, and obtain an image containing all pores and cracks. Combine the connectivity test and the aspect ratio screening criteria to filter the cracks and obtain an image containing only all cracks.

[0047] The threshold segmentation method is used to divide pores with larger apertures and cracks with larger openings, while the brightness difference segmentation method is used to divide pores with smaller apertures and cracks with smaller openings. Finally, the results of the two segmentation methods are unioned to achieve qualitative division of all pores and cracks.

[0048] The specific method of connectivity testing is: the same pore or crack is marked with the same number, and the marking numbers of pores or cracks that are not connected to each other are different.

[0049] Currently, the method for classifying cracks and pores is mainly based on morphological screening methods. By calculating the aspect ratio of the minimum circumscribed rectangle of each pore and crack, combined with the threshold screening standard of the aspect ratio, pores and cracks are divided. The lower limit of the crack aspect ratio determined by this method is 4, that is, a crack is defined as a crack if the aspect ratio is greater than 4, and a pore is defined as a pore if the aspect ratio is less than 4.

[0050] S6: removing all cracks divided in step S5 from the image of all pores and cracks in step S5 by a subtraction operation, so that the image after removal contains both organic and inorganic pores, and performing a connectivity test and numbering on all pores;

[0051] Connectivity is evaluated as follows: For a single pixel in a two-dimensional plane, it has four adjacent pixels and four diagonal pixels. If the values ​​of adjacent pixels or diagonal pixels match the pixel value of the pixel, the two pixels are considered connected. Connected pores are assigned the same number, while unconnected pores are assigned different numbers.

[0052] S7: Using a boundary expansion algorithm, the outer boundaries of all pores in step S6 are expanded outward by 1 pixel. The organic matter refined and corrected in step S4 and the pores after boundary expansion are superimposed on the same base map. The boundaries of the organic pores in the overlapping parts are constructed by taking the intersection method. The pores in the overlapping parts are screened to obtain all the numbers corresponding to the organic pores.

[0053] This step exploits the unique relationship between organic pores and organic matter to identify organic pores. Since organic pores primarily arise from the thermal evolution of organic matter, their locations are always within the organic matter. Therefore, the outer boundaries of the organic pores and the inner boundaries of the organic matter touch in the image. Expanding the outer boundaries of the organic pores will cause them to overlap with the organic matter. By selecting the pore numbers of the overlapping regions, all organic pores can be identified. However, the outer boundaries of inorganic pores generally have no correlation with the organic matter.

[0054] S8: Using the numbered image in step S6 as a sample space, only the pores corresponding to the organic pore numbers in step S7 are screened to obtain an image containing only organic pores; and organic pores are subtracted from the sample space by a subtraction operation to obtain an image containing only inorganic pores.

[0055] S9: taking a union of the organic matter image refined and corrected in step S4 and the image containing all pores and fractures in step S5, and taking a complement of the unioned images to obtain a shale matrix mineral image;

[0056] The division of shale matrix minerals is primarily performed using Boolean operations. Since the organic matter and all pores and fractures have already been divided and corrected in steps S4 and S5, the use of Boolean operations can quickly divide the shale matrix minerals, avoiding division errors caused by shadows on the pore background and slopes.

[0057] S10: In the same image, the organic matter after refinement and correction in step S4, the cracks screened out in step S5, the organic pores and inorganic pores screened out in step S8, and the shale matrix minerals in step S9 are superimposed respectively, and the above five components are marked with image numbers 0, 1, 2, 3, and 4, respectively, thereby achieving multi-component refinement and qualitative segmentation of the scanned grayscale image.

[0058] Example 1:

[0059] The shale SEM grayscale images used in this example were obtained from an organic-rich core from a deep marine shale gas reservoir in a block in the Sichuan Basin. The scanning equipment was a Zeiss SIGMA 500 field-emission scanning electron microscope. The image scanning accuracy was 10 nm, the resolution was 1024 × 1024 × 1024, the physical dimensions were 10.24 μm × 10.24 μm × 10.24 μm, and the image grayscale values ​​ranged from 0 to 255.

[0060] The following demonstrates how a multi-component segmentation method for shale scanning electron microscope images proposed in the present invention can segment the grayscale image described above, and perform multi-component superposition and verification based on the segmentation results.

[0061] The grayscale image of the shale scanning electron microscope used in this example is as follows: Figure 1 As shown, there is a large microcrack in the center of the grayscale image that extends along the edge of the organic matter. The surface roughness of the crack is large, and the tortuosity along the organic matter on the left side is large. At the same time, there are multiple microcracks with smaller openings in the upper right and lower right parts of the image. Organic matter and organic pores are developed in the figure, mainly on the left side of the image, with good connectivity. Silica minerals such as quartz and feldspar are developed on the right side of the image. At the same time, a small number of inorganic pores are developed in the matrix rock area, mainly dissolution pores, and their connectivity is poor.

[0062] The shale SEM grayscale image was preprocessed using contrast enhancement and non-local homogeneous filtering. The contrast-enhanced image was used to verify the accuracy of organic matter delineation in step S4, while noise removal using non-local homogeneous filtering improved the accuracy of pore and fracture delineation in step S5.

[0063] The initial division of organic matter uses the threshold segmentation method, and the division results are as follows: Figure 2 As shown in (a), the organic matter is over-classified due to the influence of the background of the pores and the shadow of the slope. Using image editing, the pixel value of the area over-classified as organic matter is changed from 1 to 0 under the contrast condition of the superimposed base map. The corrected secondary classification result of organic matter is shown in Figure 2 (b) shown.

[0064] In shale scanning electron microscope images, fractures and pores have the lowest grayscale range and are easily distinguished, often appearing black. A threshold segmentation method is used to segment larger pores and fractures with wide apertures. A brightness difference segmentation method is also used to identify smaller pores and fractures. Finally, the results of the two segmentation methods are combined to achieve a qualitative classification of all pores and fractures. Fractures are screened using a connectivity test and aspect ratio screening criteria to obtain an image containing only all fractures.

[0065] By using the special relationship between organic pores and organic matter in the image, the qualitative screening of organic pores is accurately achieved through the boundary expansion algorithm and the intersection method, and the inorganic pores (such as Figure 3 (a)). The position distribution of organic matter, organic pores and inorganic pores divided in this embodiment is as follows Figure 3 (b) shown.

[0066] The organic matter divided in step S4 and all the pores and cracks divided in step S5 are superimposed on the same image and the superimposed image is binarized. The binarization method here is to take the union of the organic matter, all the pores and cracks, mark the image area after the union as 1, and the rest of the non-image area is uniformly marked as 0. Take the complement of the image area after the union, mark the complement area as 1 and the rest of the area as 0 in another base image, and the shale matrix minerals divided are as follows: Figure 4 Shown in yellow.

[0067] In the same image, matrix minerals, organic matter, cracks, organic pores and inorganic pores are superimposed respectively, where shale matrix minerals are represented by yellow, organic matter is represented by green, cracks are represented by red, organic pores are represented by light blue, and inorganic pores are represented by dark blue, thus completing the multi-component qualitative segmentation of shale SEM images (such as Figure 4 The specific processing flow diagram of the present invention is shown in Figure 5 .

[0068] This paper provides a multi-component classification method for shale SEM images. This method uses the morphology and distribution characteristics of different pore types as the primary classification basis, employing digital core technology and Boolean operations as the primary mathematical methods. It employs specific classification methods for different components, theoretically demonstrating a method for classifying shale SEM grayscale images into five component images. This method is simple to operate and produces highly accurate classification results. In addition to qualitatively segmenting organic and inorganic pores, it can also be used for quantitative calculation of various pore parameters, demonstrating its broad application in shale gas reservoir evaluation and exploration and development.

[0069] The above description does not limit the present invention in any form. Although the present invention has been disclosed through the above embodiments, it is not intended to limit the present invention. Any technician familiar with the profession can make some changes or modifications to equivalent embodiments of equivalent changes using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A multi-component classification method for shale scanning electron microscope images, characterized in that: The following steps are involved: S1: Scan the polished surface of the shale sample using a nano-scanning electron microscope to obtain a two-dimensional grayscale image of the core; S2: Preprocessing the two-dimensional core grayscale image obtained in step S1 to improve image contrast and eliminate image artifacts and noise; S3: using a threshold segmentation method to perform a preliminary segmentation of the organic matter in the image pre-processed in step S2; S4: superimposing the pre-processed image in step S2 and the organic matter image initially divided in step S3, changing the transparency of either image to detect over-divided areas, and performing fine correction on the over-divided areas through image editing to obtain a secondary division image of the organic matter; S5: Combine the threshold segmentation method and the brightness difference segmentation method to segment all pores and cracks with low grayscale values, and obtain an image containing all pores and cracks. Combine the connectivity test and the aspect ratio screening criteria to filter the cracks and obtain an image containing only all cracks. S6: removing all cracks divided in step S5 from the image of all pores and cracks in step S5 by a subtraction operation, so that the image after removal contains both organic and inorganic pores, and performing a connectivity test and numbering on all pores; S7: Using a boundary expansion algorithm, the outer boundaries of all pores in step S6 are expanded outward by 1 pixel. The organic matter refined and corrected in step S4 and the pores after boundary expansion are superimposed on the same base map. The boundaries of the organic pores in the overlapping parts are constructed by taking the intersection method. The pores in the overlapping parts are screened to obtain all the numbers corresponding to the organic pores. S8: Using the numbered image in step S6 as a sample space, only the pores corresponding to the organic pore numbers in step S7 are screened to obtain an image containing only organic pores; and organic pores are subtracted from the sample space by a subtraction operation to obtain an image containing only inorganic pores. S9: taking a union of the organic matter image refined and corrected in step S4 and the image containing all pores and fractures in step S5, and taking a complement of the unioned images to obtain a shale matrix mineral image; S10: In the same image, the organic matter after refinement and correction in step S4, the cracks screened out in step S5, the organic pores and inorganic pores screened out in step S8, and the shale matrix minerals in step S9 are superimposed respectively, and the above five components are marked with image numbers 0, 1, 2, 3, and 4, respectively, thereby achieving multi-component refinement and qualitative segmentation of the scanned grayscale image.

2. The multi-component classification method of shale scanning electron microscope images according to claim 1, characterized in that: The preparation process of the polished surface of the shale sample in step S1 is as follows: first, the surface of the shale sample is rock-ground using ultra-thin emery paper until the surface is flat; then, the surface of the sample is polished using argon ion polishing equipment under vacuum conditions to obtain a smooth polished surface.

3. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: The method for improving the image contrast in step S2 is contrast enhancement. The grayscale value of the shale image after scanning electron microscopy is often dark, the image contrast is poor, and the narrow grayscale gradient will bring difficulties to subsequent image segmentation. The contrast enhancement algorithm can proportionally amplify the grayscale gradient without changing the original image information. The method for eliminating noise in step S2 is the non-local mean filtering method. Due to the presence of electronic signal interference, scanning electron microscope images generally require filtering. For the multi-component division of shale images, it is particularly important to determine the boundaries of each mineral component in the image. Using the non-local mean filtering method to filter the image can eliminate noise while preserving the image boundary information as much as possible.

4. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: In step S3, the organic matter is mainly carbonaceous components, which often appear as light gray features with medium grayscale in the scanning electron microscope image. By selecting the lower and upper grayscale thresholds of the organic matter, the positioning mark of the organic matter components can be achieved.

5. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: In order to check the accuracy of the division results in step S4, the organic matter after the initial division in step S3 and the grayscale image preprocessed in step S2 are superimposed together, and the inaccurate division areas are dynamically observed and confirmed by repeatedly changing the transparency of the organic matter image; the over-divided organic matter areas are mainly the shadows of the polished surface bevel and the background of the pores and seams. The pixel values ​​of the areas incorrectly marked as organic matter are changed from 1 to 0 through an image editor to complete the fine correction of the organic matter.

6. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: In step S5, a threshold segmentation method is used to divide pores with larger apertures and cracks with larger openings, while a brightness difference segmentation method is used to divide pores with smaller apertures and cracks with smaller openings. Finally, the results of the two segmentation methods are combined to achieve a qualitative division of all pores and cracks.

7. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: The connectivity test method in step S5 is specifically as follows: the same pore or crack is marked with the same number, and the pores or cracks that are not connected to each other are marked with different numbers; At present, the method of dividing cracks and pores is based on the morphological screening method. The aspect ratio of the minimum circumscribed rectangle of each pore and crack is calculated, and the threshold screening standard of the aspect ratio is combined to divide pores and cracks. The lower limit of the determined crack aspect ratio is 4, that is, the aspect ratio is greater than 4 and it is a crack, and the aspect ratio is less than 4 and it is a pore.

8. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: The evaluation criteria for connectivity in step S6 are as follows: for a single pixel point in a two-dimensional plane, it has 4 adjacent pixels and 4 diagonal pixels. When the adjacent pixel value or the diagonal pixel value is consistent with the pixel value of the pixel point, the two pixels are considered to be connected; interconnected pores are marked with the same number, and unconnected pores are marked with different numbers.

9. The multi-component classification method of shale scanning electron microscope images as claimed in claim 1, characterized in that: In step S7, the special relationship between organic pores and organic matter is used to screen organic pores. Since organic pores are mainly formed by thermal evolution of organic matter, the space where the organic pores are located is always inside the organic matter. Therefore, the outer boundary of the organic pore and the inner boundary of the organic matter are in contact with each other in the image. After the outer boundary of the organic pore is expanded outward, it will overlap with the organic matter. By screening the pore numbers of the overlapping part, the numbers of all organic pores can be obtained, while the outer boundary of the inorganic pore has no correlation with the organic matter.

10. The multi-component classification method of shale scanning electron microscope images according to claim 1, characterized in that: In step S9, Boolean operations are used to divide the shale matrix minerals. Since the organic matter and all pores and fractures have been divided and corrected in steps S4 and S5, the shale matrix minerals can be quickly divided using Boolean operations, avoiding division errors caused by pore backgrounds and slope shadows.

Images