A method for generating a particle structure map based on rock specimens and related equipment

Abstract

The present application provides a method for generating a particle structure atlas based on rock specimens and related equipment, which relates to the field of graphic data reading technology. A cross-sectional image sequence is obtained by three-dimensional scanning to obtain a cross-sectional image of the internal structure of the rock specimen, and then the seed point is determined based on the grayscale value of the image. These seed points serve as the starting point as the initial voxel. Since different particles have different image grayscale values, the seed point is used as the basis, and by comparing the grayscale values ​​with adjacent voxels and classifying adjacent voxels that meet the conditions as the same particle according to the preset growth threshold conditions, this process effectively simplifies the complexity of particle aggregation. It avoids many interference factors encountered when the entire image is processed indiscriminately, and improves the efficiency of generating structural atlases.

Description

Technical Field

[0001] The present application relates to the technical field of graphic data reading, and in particular to a method for generating a particle structure map based on rock specimens and related equipment. Background Art

[0002] In geological research and engineering applications, a thorough analysis of the particle structure of rock specimens is crucial for in-depth investigation of rock physical and chemical properties and prediction of rock behavior under varying stress conditions. Accurate particle structure analysis can provide strong support for mine planning and geological disaster prevention.

[0003] The relevant technology uses electron microscopy image analysis and 3D reconstruction techniques to analyze rock specimens. Electron microscopes perform high-resolution imaging of rock specimens, acquiring numerous microscopic images. These images are then processed using image analysis algorithms to extract information such as particle boundaries and shapes. 3D reconstruction algorithms then integrate this information into a 3D model of the rock's grain structure, which serves as an atlas.

[0004] However, the amount of computation required to integrate information such as boundaries and shapes is too large, resulting in low efficiency in generating structural maps. Summary of the Invention

[0005] The present application provides a method for generating a particle structure map based on rock specimens and related equipment, which are used to improve the efficiency of map generation.

[0006] In a first aspect, the present application provides a method for generating a particle structure atlas based on a rock specimen, comprising: performing a three-dimensional scan of the rock specimen to obtain a cross-sectional image sequence of the internal structure of the rock specimen; identifying a seed point from the cross-sectional image sequence based on the image grayscale value, and setting the seed point as an initial voxel; performing a grayscale value comparison on adjacent voxels of the initial voxel, and assigning adjacent voxels that meet a preset growth threshold condition to the same particle; performing a feature analysis on all acquired particles to obtain digital features, the digital features including: morphological characteristic parameters of the particles, spatial distribution characteristics of the particles, a contact relationship network between particles, and a void network between particles; projecting the particles onto a two-dimensional plane according to the digital features to obtain a projection image; generating an analysis report and a visualization atlas, wherein the analysis report includes the morphological characteristic parameters of the particles, the spatial distribution characteristics of the particles, the contact relationship network between particles, and the void network between particles, and the visualization atlas includes the projection image.

[0007] By adopting the above technical solution, a sequence of cross-sectional images is obtained using three-dimensional scanning to obtain cross-sectional images of the internal structure of the rock specimen. Then, seed points are determined based on the image grayscale values. These seed points serve as the starting point for the initial voxels. Since different particles have different image grayscale values, the seed points are used as the basis, and by comparing the grayscale values ​​of adjacent voxels and classifying adjacent voxels that meet the conditions as the same particle based on the preset growth threshold conditions, this process effectively simplifies the complexity of particle aggregation. It avoids the many interference factors faced when the entire image is processed indiscriminately, and improves the efficiency of generating structural maps. In addition, all particles are digitized, and the information of the physical particles is successfully converted into digital form, which facilitates the precise generation of subsequent maps and makes the entire particle structure analysis process more efficient.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of identifying a seed point from a cross-sectional image sequence based on an image grayscale value and setting the seed point as an initial voxel specifically includes: using the grayscale value in the cross-sectional image sequence as a height value to construct a three-dimensional height field, wherein the position of each voxel point is determined by the corresponding spatial coordinate, and the height value of the voxel point is determined by the corresponding grayscale value; sorting the voxel points in the three-dimensional height field in ascending order; adding the voxel points as vertices to a simplex complex in the order after the ascending order, and constructing each voxel point into a zero-dimensional simplex structure; in the three-dimensional height field, Adjacent voxel points of a voxel point are searched within a preset neighborhood range, and two voxel points whose distance is less than a distance threshold are connected to construct a one-dimensional simplex; when the endpoints of the one-dimensional simplex edge are connected end to end to form a closed loop, the area enclosed by the closed loop is constructed as a two-dimensional simplex surface, and the two-dimensional simplex surface is added to the simplex complex; a nested simplex complex sequence is formed, in which each layer of the simplex complex sequence represents a topological structure with a grayscale value less than or equal to the grayscale value of the current voxel point; the persistence of the nested simplex complex sequence is calculated, and the local minimum point whose persistence is greater than the persistence threshold is calculated as the seed point.

[0009] By adopting the above technical solution, the grayscale values ​​in the cross-sectional image sequence are converted into the height values ​​of the three-dimensional height field, and the corresponding relationship between the voxel point position and the height value is determined, laying the foundation for the subsequent topological structure construction. Sorting the voxel points in ascending order facilitates orderly processing, and constructing zero-dimensional, one-dimensional, and two-dimensional simplex structures in turn to form a nested simplex complex sequence. This process integrates the spatial and grayscale information system of the voxel points into topological structure information. By performing persistence calculations on the nested simplex complex sequence and using local minimum points as seed points, it is possible to accurately screen out voxel points that are important in the evolution of the topological structure, eliminate a large amount of irrelevant information interference, optimize the accuracy and efficiency of seed point selection, and make subsequent seed point-based particle recognition and analysis more accurate and reliable.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the step of performing persistence calculation on the nested simplex-complex sequence specifically includes: a new connected component is generated at a local minimum point, and the local minimum point is recorded as the birth point of the connected component; the connected component is formed by the interconnection of voxel points within the nested simplex-complex sequence, and the local minimum point is the point with the smallest grayscale value in the pre-divided local area; when two connected components are connected and merged for the first time at the current grayscale value, the connected component with the higher birth value is terminated, and the current grayscale value is recorded as the extinction point corresponding to the connected component with the higher birth value, and the birth value refers to the grayscale value corresponding to the birth point of the connected component; a persistence graph is constructed based on the birth point and the corresponding extinction point, wherein the horizontal axis of the persistence graph is the grayscale value of the birth point, and the vertical axis of the persistence graph is the grayscale difference between the birth point and the corresponding extinction point.

[0011] By employing this technical solution, during the connected component analysis process, the birth point of the connected component is determined by the local minimum point. Connected components are formed based on voxel connectivity. When connections merge, the extinction point is determined based on the birth value and a persistence graph is constructed. This method, based on recording key nodes of topological evolution, deeply transforms image grayscale information into a quantitative description of topological structural features. The persistence graph intuitively displays the existence range and stability of connected components. By utilizing information related to the birth and extinction points, it can effectively evaluate the topological changes of the particle structure under different grayscale value changes, making the selection of seed points more accurate.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the step of performing feature analysis on all the acquired particles to obtain digital features specifically includes: calculating the ratio of the surface area to the volume of each particle to obtain the sphericity parameter; calculating the ratio of the longest axis to the shortest axis of each particle to obtain the elongation parameter.

[0013] By employing this technical solution, we can calculate the particle's sphericity and elongation parameters. The sphericity parameter reflects the particle's sphericity by measuring the surface area to volume ratio, while the elongation parameter, measuring the ratio of the longest axis to the shortest axis, reflects the particle's elongation or flattening. These two key parameters succinctly and effectively quantify the particle's morphological characteristics, significantly reducing data processing and computational complexity compared to complex full-shape information processing.

[0014] In combination with some embodiments of the first aspect, in some embodiments, the step of performing feature analysis on all the acquired particles to obtain digital features specifically includes: statistically calculating the particle size distribution of the particles; constructing a position density distribution map of the particles in three-dimensional space based on the coordinate information of the particles in three-dimensional space to obtain the spatial distribution characteristics of the particles.

[0015] By adopting the above technical solution, the particle size distribution is statistically analyzed and a three-dimensional spatial position density distribution map is constructed. The particle size distribution intuitively presents the composition of particle sizes. Particles of different sizes play different roles in the mechanical structure of rocks. Its statistical results provide key data for evaluating the overall particle size characteristics and mechanical properties of rocks. The position density distribution map constructed based on the three-dimensional spatial coordinates of the particles clearly shows the sparse and dense layout of particles in space, which helps to further study the uniformity and local differences of the internal structure of the rock. By obtaining these two features, the spatial distribution overview of particles can be efficiently grasped without in-depth analysis of the complex shape boundaries of the particles, reducing the data processing steps.

[0016] In combination with some embodiments of the first aspect, in some embodiments, feature analysis is performed on all acquired particles to obtain digital features, specifically including: identifying contact areas between adjacent particles; calculating the area and perimeter of the contact areas; classifying the contact areas as point contact, line contact, or surface contact based on the area and perimeter; and constructing a contact relationship network between particles based on the contact relationship between particles.

[0017] By employing this technical solution, contact areas between adjacent particles are identified and multi-dimensionally analyzed to construct a contact relationship network. By precisely locating contact areas, calculating their area and perimeter, and classifying them into different contact types, each of which exhibits significant differences in force transmission and intensity. The network constructed based on these contact relationships intuitively displays the topological structure of inter-particle connections. Compared to fully integrating all particle information, focusing on contact relationships reduces non-critical information processing and provides clear and explicit relationships between particle interactions.

[0018] In combination with some embodiments of the first aspect, in some embodiments, the steps of performing feature analysis on all acquired particles to obtain digital features specifically include: identifying the void areas between particles; statistically analyzing the size distribution of the void areas; and constructing a void network between particles.

[0019] By employing these technical solutions, we identify intergranular voids and calculate their size distribution to construct a void network. The shape, size, and distribution of voids critically influence the integrity and functionality of the rock's internal structure. Statistical size distribution accurately quantifies the proportion and distribution of voids of different sizes, while the constructed void network systematically identifies the spatial correlations and connectivity between voids. This precise data on void areas and network structure information can be incorporated into the final analysis report and visualization.

[0020] In a second aspect, the present application provides a system for generating a particle structure map based on a rock specimen, and the system for generating a particle structure map based on a rock specimen includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, and the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the system for generating a particle structure map based on a rock specimen to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, the present application provides a computer program product comprising instructions. When the computer program product is run on a particle structure map generation system based on rock specimens, the particle structure map generation system based on rock specimens executes the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a particle structure map generation system based on rock specimens, enables the particle structure map generation system based on rock specimens to execute the method described in the first aspect and any possible implementation of the first aspect.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0024] 1. Use 3D scanning to obtain a sequence of cross-sectional images to obtain cross-sectional images of the internal structure of the rock specimen. Then, seed points are determined based on the image grayscale values. These seed points serve as the starting point for the initial voxels. Since different particles have different image grayscale values, the seed points are used as the basis. By comparing the grayscale values ​​of adjacent voxels and classifying adjacent voxels that meet the conditions as the same particle based on the preset growth threshold conditions, this process effectively simplifies the complexity of particle aggregation. It avoids the many interference factors faced when the entire image is processed indiscriminately and improves the efficiency of generating structural maps. In addition, all particles are digitized and the information of the physical particles is successfully converted into digital form, which facilitates the precise generation of subsequent maps and makes the entire particle structure analysis process more efficient.

[0025] 2. Convert the grayscale values ​​in the cross-sectional image sequence into height values ​​of the three-dimensional height field, determine the correspondence between the voxel point position and the height value, and lay the foundation for the subsequent topological structure construction. Sorting the voxel points in ascending order facilitates orderly processing, and constructing zero-dimensional, one-dimensional, and two-dimensional simplex structures in turn to form a nested simplex complex sequence. This process integrates the spatial and grayscale information systems of the voxel points into topological structure information. By performing persistence calculations on the nested simplex complex sequence and using local minimum points as seed points, it is possible to accurately screen out voxel points that are important in the evolution of the topological structure, eliminate a large amount of irrelevant information interference, optimize the accuracy and efficiency of seed point selection, and make subsequent seed point-based particle recognition and analysis more accurate and reliable.

[0026] 3. During the connected component analysis process, local minima are used to determine the birth points of connected components. Connected components are formed based on voxel connectivity. When connections merge, the extinction points are determined based on the birth values, and a persistence graph is constructed. This method, based on recording key nodes of topological evolution, deeply transforms image grayscale information into a quantitative description of topological structural features. The persistence graph intuitively demonstrates the existence range and stability of connected components. By utilizing information related to the birth and extinction points, it can effectively evaluate the topological changes of the particle structure under different grayscale value changes, making the selection of seed points more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a method for generating a particle structure map based on rock specimens in an embodiment of the present application;

[0028] Figure 2 is another flow chart of a method for generating a particle structure map based on a rock specimen in an embodiment of the present application;

[0029] Figure 3 yes Figure 2 A schematic diagram of a specific flow of sub-step S2071 of step S207;

[0030] Figure 4 This is a schematic diagram of an exemplary hardware structure of a particle structure atlas generation system based on rock specimens in an embodiment of the present application. DETAILED DESCRIPTION

[0031] The terms used in the following examples of this application are for the purpose of describing specific embodiments only and are not intended to limit the application. As used in the specification and appended claims of this application, the singular expressions "a," "an," "said," "the," and "this" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the terms "and / or" used in this application refer to and encompass any and all possible combinations of one or more of the listed items.

[0032] In the following, the terms first and second are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of technical features indicated. Therefore, features defined as first or second may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, unless otherwise specified, the meaning of "more than" is two or more.

[0033] See also Figure 1 , Figure 1 This is a flow chart of a method for generating a particle structure map based on rock specimens in an embodiment of the present application;

[0034] S101, performing three-dimensional scanning on a rock specimen to obtain a sequence of cross-sectional images of the internal structure of the rock specimen;

[0035] 3D scanning refers to the use of specific 3D scanning equipment to collect non-contact data from multiple angles on rock specimens, in order to obtain information about the specimen's shape and structure in three dimensions. A cross-sectional image sequence is a collection of 2D images corresponding to each layer of a rock specimen, virtually layered at a specific thickness or interval. These images are arranged in sequence to represent structural information at different depths within the rock specimen.

[0036] The rock specimens are comprehensively scanned using 3D scanning equipment within a preset accuracy and scanning range. During the scanning process, the equipment records a large amount of spatial coordinate data and corresponding information such as the intensity of reflected light. Specialized image processing software is then used to convert this data into a series of cross-sectional image sequences. These images can clearly show structural details such as the texture and particle distribution at different levels within the rock specimen.

[0037] S102, identifying a seed point from the cross-sectional image sequence based on the image grayscale value, and setting the seed point as an initial voxel;

[0038] The image grayscale value refers to the brightness information of each pixel in the image, expressed as a numerical value and used to distinguish the brightness and darkness of different areas in the image. It reflects the differences in light reflection or absorption of different materials or structures within the rock specimen. In image analysis, the seed point is the starting reference point used to initiate subsequent specific region identification or growth algorithms. In this step, it is the key point used to indicate the starting position of particle identification. The initial voxel is an extension of the seed point concept in three-dimensional space. It refers to the smallest unit of a certain volume constructed based on the seed point in three-dimensional image data, similar to a three-dimensional pixel.

[0039] Specifically, after acquiring a sequence of cross-sectional images, to accurately identify the grain structure in a rock specimen, it is necessary to determine the starting point for particle identification. This step involves performing a grayscale analysis on each image in the cross-sectional image sequence. Because different particles may exhibit different grayscale characteristics in the image, a specific grayscale value range or grayscale variation pattern is set to select pixels that meet the requirements as seed points. These seed points are then defined as initial voxels, providing a foundation for subsequent particle growth and identification.

[0040] In some specific embodiments, the grayscale normalization processing is first performed on the cross-sectional image sequence to make the grayscale value range of all images uniform, and then the grayscale histogram of each image is calculated. The possible grayscale range of the particle is determined based on the peak and valley values ​​of the histogram. Within this range, pixel points with relatively stable grayscale values ​​and a high probability of being at the center of the particle are selected as seed points and converted into initial voxels, which are not limited here.

[0041] S103, comparing the grayscale values ​​of adjacent voxels of the initial voxel, and assigning adjacent voxels that meet a preset growth threshold condition to the same particle;

[0042] The preset growth threshold condition refers to a pre-set restriction condition on the difference range of the grayscale value between adjacent voxels and the initial voxel, which is used to determine whether adjacent voxels belong to the same particle structure.

[0043] After determining the initial voxel, this step is necessary to fully identify the range of the particle. By traversing the neighboring voxels around the initial voxel and comparing their grayscale values ​​with the initial voxel one by one, if the difference is within the preset growth threshold, the neighboring voxel is considered to belong to the same particle structure as the initial voxel and is assigned to the particle. This process continues outward to gradually determine the range of the entire particle.

[0044] In some specific embodiments, a three-dimensional voxel traversal queue is first established, an initial voxel is placed in the queue, and then a voxel is taken out of the queue, and the grayscale value differences between all its adjacent voxels and it are calculated. Adjacent voxels that meet the preset growth threshold conditions are added to the queue and marked as belonging to the same particle. This process is repeated until no new voxels are added to the queue, which is not limited here.

[0045] S104, performing feature analysis on all acquired particles to obtain digital features, including: morphological characteristic parameters of the particles, spatial distribution characteristics of the particles, contact relationship networks between particles, and gap networks between particles;

[0046] Among them, the morphological characteristic parameters of particles refer to numerical indicators used to quantitatively describe the shape characteristics of individual particles, such as sphericity, elongation, etc., which reflect the external contour characteristics of the particles.

[0047] Specifically, after all particles are identified, a comprehensive characterization of these particles is required to gain a deeper understanding of the rock specimen's grain structure. This step is then performed. A specialized algorithm is used to calculate the morphological parameters of each particle, such as sphericity by the surface area to volume ratio and elongation by the ratio of the longest axis to the shortest axis. The spatial distribution of particles is statistically analyzed based on their coordinates in three-dimensional space, such as the number of particles per unit volume to determine density. A contact relationship network is constructed by identifying contact areas between adjacent particles and analyzing their characteristics. Finally, the void areas between particles are identified and their related information is statistically analyzed to construct a void network, ultimately yielding these digital features.

[0048] In some embodiments, the ratio of the surface area to the volume of each particle is calculated to obtain the sphericity parameter; and the ratio of the longest axis to the shortest axis of each particle is calculated to obtain the elongation parameter.

[0049] As can be seen, the sphericity and elongation parameters of the particles are calculated. The sphericity parameter reflects the degree of particle sphericity through the ratio of surface area to volume, while the elongation parameter reflects the elongation or flattening of the particle through the ratio of the longest axis to the shortest axis. These two key parameters succinctly and effectively quantify the morphological characteristics of the particles, significantly reducing the amount of data processing and computational complexity compared to complex full shape information processing.

[0050] In some embodiments, the particle size distribution of the particles is statistically analyzed; based on the coordinate information of the particles in the three-dimensional space, a position density distribution diagram of the particles in the three-dimensional space is constructed to obtain the spatial distribution characteristics of the particles.

[0051] As can be seen, the particle size distribution is statistically analyzed and a three-dimensional spatial position density distribution map is constructed. The particle size distribution intuitively presents the composition of particle sizes. Particles of different sizes play different roles in the mechanical structure of rocks. Its statistical results provide key data for evaluating the overall particle size characteristics and mechanical properties of rocks. The position density distribution map constructed based on the three-dimensional spatial coordinates of the particles clearly shows the sparse and dense layout of particles in space, which helps to further study the uniformity and local differences of the internal structure of the rock. By obtaining these two features, the spatial distribution of particles can be efficiently grasped without in-depth analysis of the complex shape boundaries of the particles, reducing the data processing steps.

[0052] In some embodiments, contact regions between adjacent particles are identified; the area and perimeter of the contact regions are calculated; the contact regions are classified as point contacts, line contacts, or surface contacts based on the area and perimeter; and a contact relationship network between particles is constructed based on the contact relationship between particles.

[0053] As can be seen, the contact areas between adjacent particles are identified and analyzed in multiple dimensions to construct a contact relationship network. By precisely locating contact areas, calculating their area and perimeter, and classifying them into different contact types, different contact types exhibit significant differences in force transmission and intensity. The network constructed based on these contact relationships intuitively displays the topological structure of inter-particle connections. Compared to fully integrating all particle information, focusing on contact relationships reduces non-critical information processing and makes the interactions between particles clear and explicit.

[0054] In some embodiments, void regions between particles are identified; size distribution of the void regions is calculated; and a void network between particles is constructed.

[0055] As can be seen, identifying interparticle voids and calculating their size distribution to construct a void network. The shape, size, and distribution of voids critically influence the integrity and functionality of the rock's internal structure. Statistical size distribution accurately quantifies the proportion and distribution patterns of voids of different sizes, while the constructed void network systematically identifies the spatial correlations and connectivity between voids. This precise data on void areas and network structure information can be incorporated into the final analysis report and visualization.

[0056] S105, projecting the particles onto a two-dimensional plane according to the digital features to obtain a projection image;

[0057] After acquiring the digital features of the particles, this step is necessary to present the particle structure in a more intuitive manner for easier observation and analysis. Using a pre-defined projection algorithm, such as orthogonal or perspective projection, the particles in three-dimensional space are projected onto a two-dimensional plane based on their digital features (e.g., coordinates and shape). While the particles may be deformed during the projection process, their relative positional relationships and key features are preserved, resulting in a projected image.

[0058] In some specific embodiments, the projection type is determined to be orthogonal projection, and then the size and position of the two-dimensional projection plane are set according to the coordinate range of the particles in three-dimensional space, and the three-dimensional coordinates of each particle are converted into two-dimensional coordinates. The shape of the particles is simplified according to their morphological characteristic parameters, such as approximating a complex shape to an ellipse or rectangle and then projecting it onto a two-dimensional plane. This is not limited here.

[0059] In some specific embodiments, a perspective projection method is used to first determine the viewpoint position and projection direction of the perspective projection, and then calculate the projection coordinates of each particle based on the perspective principle. During the projection process, the projection size is adjusted according to the distance between the particle and the viewpoint to reflect the perspective effect. At the same time, the contact relationship and void network of the particles and other features are marked on the projected image with specific symbols or colors, which are not limited here.

[0060] S106. Generate an analysis report and a visualization map, wherein the analysis report includes morphological characteristic parameters of the particles, spatial distribution characteristics of the particles, contact relationship network between particles, and gap network between particles, and the visualization map includes a projection image.

[0061] After completing the previous steps of particle identification, feature analysis, and projection, this step is performed to comprehensively summarize and present the research results. The particle morphological parameters are organized into a data table, the spatial distribution characteristics are described in text, and the structure and characteristics of the contact relationship network and void network are elaborated in detail to generate an analysis report. Simultaneously, the projected image is typeset, with necessary annotations and legends added. It is then combined with other possible visualization elements (such as a particle feature distribution histogram) to generate a visualization map.

[0062] It can be seen that by using 3D scanning to obtain a sequence of cross-sectional images, a cross-sectional image of the internal structure of the rock specimen is obtained. Then, seed points are determined based on the image grayscale values. These seed points serve as the starting point for the initial voxels. Since different particles have different image grayscale values, this process effectively simplifies the complexity of particle aggregation by comparing the grayscale values ​​of adjacent voxels with those of adjacent voxels and classifying adjacent voxels that meet the conditions as the same particle based on the preset growth threshold conditions. It avoids the many interference factors faced when the entire image is processed indiscriminately and improves the efficiency of generating structural maps. In addition, all particles are digitized and the information of the physical particles is successfully converted into digital form, which facilitates the precise generation of subsequent maps and makes the entire particle structure analysis process more efficient.

[0063] In actual use, relying solely on grayscale values ​​can lead to a problem: because different types of particles have different grayscale values, the selection of seed points can be overly biased and unstable. Because the composition of various particles in rock specimens is complex and diverse, the distribution ranges of their grayscale values ​​may overlap or intersect. Selecting seed points based solely on grayscale thresholds can easily misjudge or miss the starting points for some particles, leading to significant deviations in subsequent particle identification and analysis based on these inaccurate seed points.

[0064] See also Figure 2 , Figure 2 is another flow chart of a method for generating a particle structure map based on a rock specimen in an embodiment of the present application;

[0065] In some embodiments, step S102 specifically includes:

[0066] S201, using the grayscale values ​​in the cross-sectional image sequence as height values ​​to construct a three-dimensional height field, wherein the position of each voxel point is determined by the corresponding spatial coordinates, and the height value of the voxel point is determined by the corresponding grayscale value;

[0067] Specifically, after obtaining the cross-sectional image sequence, the grayscale value of each pixel in the image is extracted and converted into coordinates in three-dimensional space according to the position of the image in the sequence and the coordinate position of the pixel in the image (the x and y coordinates correspond to the image plane position, and the z coordinate corresponds to the layer position of the image in the sequence). At the same time, the grayscale value of the pixel is assigned to the voxel point corresponding to the coordinate position as its height value. In this way, a three-dimensional height field is constructed, which integrates the original two-dimensional image grayscale information into the three-dimensional spatial structure, facilitating subsequent more in-depth topological structure analysis.

[0068] S202, sorting the voxel points in the three-dimensional height field in ascending order;

[0069] Specifically, after constructing the 3D height field, this step is necessary to orderly utilize voxel points to construct topological structures such as simplices and complexes. Using a specific sorting algorithm, such as quick sort or bubble sort, all voxel points in the 3D height field are traversed and compared based on their height values, rearranging them in ascending order of height value. This allows subsequent operations to process the voxel points in this order, gradually constructing a topological structure that meets the requirements.

[0070] S203, adding the voxel points as vertices to the simplex complex in ascending order, and constructing each voxel point into a zero-dimensional simplex structure;

[0071] The simplex complex is a topological concept used to describe spatial structures. It is composed of multiple simplices arranged according to certain rules. It embodies the connections and topological relationships between elements in space and can be seen as an abstract representation of complex spatial structures. The zero-dimensional simplex structure is the most basic and simplest form of the simplex complex. It is simply a single point. Specifically, each voxel point is considered a zero-dimensional simplex, serving as the fundamental unit for constructing more complex topological structures.

[0072] Specifically, after completing the ascending sorting of the voxels, this step is required to gradually construct a simplex complex that reflects the topological structure associated with the image's grayscale values. Voxels are extracted one by one in the sorted order and added as vertices to the simplex complex. Each added voxel is equivalent to constructing an independent zero-dimensional simplex structure within the simplex complex. These zero-dimensional simplex structures are then further developed into more complex topological structures through specific connection rules.

[0073] S204, searching for adjacent voxel points of the voxel point within a preset neighborhood of the three-dimensional height field, and connecting two voxel points whose distance is less than a distance threshold to construct a one-dimensional simplex;

[0074] Among them, the preset neighborhood range refers to a local spatial area pre-set in the three-dimensional height field, which is used to define the range of searching for adjacent voxel points of the voxel point. Its size and shape can be set according to actual needs and image characteristics. For example, it can be set to a cube or spherical space area centered on the voxel point. Adjacent voxel points refer to other voxel points that are relatively close to the target voxel point in spatial position within the preset neighborhood range, and there is a certain spatial correlation between them. A one-dimensional simplex is a topological structure that is one level more complex than a zero-dimensional simplex structure. In topology, it represents a line segment. Here, it is a line segment structure formed by connecting two voxel points that meet the distance conditions, which is used to further enrich the topological relationship represented by the simplex complex.

[0075] Specifically, after constructing the zero-dimensional simplex structure, this step is performed in order to allow the simplex complex to more carefully reflect the spatial connection relationship between voxel points and thus mine the potential topological structure information in the image. For each voxel point, with it as the center, other voxel points are searched within a preset neighborhood range, and the spatial distance between the voxel point and the adjacent voxel points found is calculated (for example, it can be measured by calculating the Euclidean distance of the difference between the coordinates of the two points). The adjacent voxel points with a distance less than the distance threshold are connected to the current voxel point, thus constructing a one-dimensional simplex between them, that is, forming a line segment-like topological connection, so that the simplex complex develops from isolated points (zero-dimensional simplex) to a structure with a connection relationship, gradually constructing a more complex structure that can reflect the grayscale-related topological characteristics of the image.

[0076] S205. When the endpoints of the one-dimensional simplex edges are connected end to end to form a closed loop, construct the region enclosed by the closed loop as a two-dimensional simplex surface, and add the two-dimensional simplex surface to the simplex complex;

[0077] A closed loop is a closed line formed by connecting multiple one-dimensional simplex edges end to end, like a circle formed by line segments. In topology, a two-dimensional simplex face is equivalent to a triangle or a polygonal plane region composed of multiple triangles. Here, it refers to the region enclosed by a closed loop. It is a planar topological structure abstracted from the construction of one-dimensional simplexes. It is used to more fully reflect the connections and spatial relationships between voxel points on a two-dimensional level and is an important component in constructing more complex topological structures.

[0078] Specifically, based on the constructed one-dimensional simplex, as the one-dimensional simplex edges continue to connect, when the endpoints are connected end to end to form a closed loop, this step is required to further improve the topological structure represented by the simplex complex so that it can more comprehensively reflect the spatial structural information related to the image's grayscale values. By detecting the connection between the one-dimensional simplex edges, once a closed loop is formed, the area enclosed by this closed loop is identified and constructed as a two-dimensional simplex surface. This two-dimensional simplex surface is then added to the simplex complex, allowing the simplex complex to evolve from a structure containing line segments (one-dimensional simplex) to a structure containing planar regions (two-dimensional simplex surfaces), continuously enriching its topological level, allowing for subsequent deeper exploration of the image's inherent topological features.

[0079] S206, forming a nested simplex complex sequence, wherein each layer of the simplex complex sequence represents a topological structure whose grayscale value is less than or equal to the grayscale value of the current voxel point;

[0080] Among them, the nested simplex complex sequence refers to a structure composed of multiple simplex complexes nested layer by layer according to certain rules. Each layer of simplex complex is associated with a specific grayscale value range and contains simplices of different dimensions from zero dimension to two dimension.

[0081] Specifically, in the process of continuously adding two-dimensional simplex faces to the simplex complex and gradually enriching its topological structure, this step is required to comprehensively and systematically present the topological structure evolution process corresponding to the change of image grayscale values ​​from low to high. As voxel points are sequentially sorted in ascending order to participate in the construction of the simplex complex, each new voxel point is determined according to its grayscale value. The currently constructed simplex complex composed of voxel points with grayscale values ​​less than or equal to the grayscale value of the voxel point is placed as this layer in the nested simplex complex sequence. This continuous accumulation forms a hierarchical nested simplex complex sequence, which can clearly show the topological structure characteristics of the image at different grayscale value stages and the relationship between them, providing a comprehensive topological structure foundation for subsequent persistence calculation and other analyses.

[0082] In some specific embodiments, an empty list is created to store a nested sequence of simplex complexes. Each voxel is then processed in ascending order. For each voxel, all simplices in the currently constructed simplex complex consisting of voxels whose grayscale values ​​are less than or equal to the grayscale value of the voxel are filtered out based on its grayscale value. These simplices are combined into a new simplex complex as a layer, which is added to the list to form a nested structure. This is not limited to this.

[0083] S207 , performing persistence calculation on the nested simplex-complex sequence, and calculating a local minimum point whose persistence is greater than a persistence threshold as a seed point.

[0084] Persistence calculation is a computational method based on topological structure evolution analysis. It is primarily used to measure the stability and duration of different topological features (such as connected components) during a certain change process (here, with the change of image grayscale values). This quantitatively reflects some key characteristics of the topological structure and helps identify points of significance in the overall image structure analysis. The persistence threshold is a pre-set numerical limit used to screen local minimum points that are sufficiently stable in the persistence calculation and worthy of being used as the starting point for subsequent particle identification.

[0085] Specifically, after forming a nested simplex complex sequence, this step is required to screen out key seed points from the complex topological structure that can serve as the basis for initial particle identification. Using a specific persistence calculation algorithm, the nested simplex complex sequence is analyzed, and the persistence characteristics of the topological structure in each local area as the grayscale value changes are calculated. The obtained persistence value is compared with a pre-set persistence threshold, and the local minimum points with a persistence greater than the threshold are selected. These points are identified as seed points that meet the requirements. They have relatively stable and important characteristics in terms of the image's topological structure and grayscale value distribution. Starting with these seed points, particles can be identified and divided more accurately in the future, thereby conducting in-depth analysis of the rock specimen's grain structure.

[0086] It can be seen that converting the grayscale values ​​in the cross-sectional image sequence into the height values ​​of the three-dimensional height field and determining the correspondence between the voxel point position and the height value lays the foundation for the subsequent topological structure construction. Sorting the voxel points in ascending order facilitates orderly processing, and constructing zero-dimensional, one-dimensional, and two-dimensional simplex structures in turn to form a nested simplex complex sequence. This process integrates the spatial and grayscale information of the voxel points into topological structure information. By performing persistence calculations on the nested simplex complex sequence and using local minimum points as seed points, it is possible to accurately screen out voxel points that are important in the evolution of the topological structure, eliminate a large amount of irrelevant information interference, optimize the accuracy and efficiency of seed point selection, and make subsequent seed point-based particle recognition and analysis more accurate and reliable.

[0087] It should be noted that the grayscale values ​​themselves do not change during this process. However, when constructing the nested simplex-complex sequence and performing persistence calculations, the topological structure is gradually constructed and its evolution is observed based on the grayscale value order of the voxel points (in ascending order).

[0088] When we say that each layer of the simplex complex sequence represents a topological structure with a grayscale value less than or equal to the grayscale value of the current voxel, it does not mean that the grayscale value has changed. Rather, during the construction process, as the voxels are processed in order from small to large grayscale values, the grayscale value of each new voxel is like a new stage. For example, initially, the voxel with the smallest grayscale value is processed to construct a simple topological structure, which is the first layer. Then, when processing the next voxel with a slightly larger grayscale value, the topological structure composed of all the previous voxels with grayscale values ​​less than or equal to the grayscale value of this new voxel is used as a new layer. This is like building a topological structure that can reflect different grayscale value stages layer by layer along the grayscale value ladder (the grayscale values ​​of the voxels after sorting in ascending order), making it easy to observe how the topological structure changes under different grayscale ranges.

[0089] See also Figure 3 , Figure 3 yes Figure 2 A schematic diagram of a specific flow of sub-step S2071 of step S207;

[0090] In some specific embodiments, step S207 specifically includes:

[0091] S2071. A new connected component is generated at a local minimum point, which is recorded as the birth point of the connected component. The connected component is formed by interconnected voxel points within the nested simplex-complex sequence. The local minimum point is the point with the smallest grayscale value within the pre-divided local area.

[0092] A connected component is an independent part of a topological structure composed of interconnected voxels, much like a set of nodes in a complex network that are connected by paths and can be considered a whole. In this case, it is a relatively independent region formed by the connections between voxels in a nested sequence of simplexes and complexes. A local minimum is a point within a pre-defined local region whose grayscale value is the smallest compared to the grayscale values ​​of other points within that region. For example, if the entire image is divided into multiple small regions, a voxel in a certain region has a lower grayscale value than all other voxels in the region. This is the local minimum point in that region. A birth point is used to indicate the starting point of the formation of a connected component. Marking a local minimum point as the birth point of a connected component means that the connected component is gradually constructed from the location of this point with the smallest grayscale value.

[0093] Specifically, within the constructed nested simplicial-complex sequence, local regions are divided according to pre-set rules (for example, by a certain spatial range or number of voxels). Within each local region, the point with the smallest grayscale value, also known as the local minimum, is searched. When voxels are found to be interconnected around this local minimum, forming a relatively independent region, this region becomes a connected component, and the local minimum is recorded as the birth point of this connected component. It marks the starting state of this connected component during the topological evolution process, and this birth point can be used to track the subsequent development and changes of the connected component.

[0094] In some specific embodiments, the space corresponding to the entire three-dimensional height field is divided into units of fixed-side-length cubes, each of which is a local region. The voxel points within each local region are then traversed, and by comparing their grayscale values, the voxel point with the smallest grayscale value is found as the local minimum point. Next, starting from this local minimum point, a graph search algorithm (such as a breadth-first search or a depth-first search algorithm) is used to search for other voxel points connected to it. The set of these interconnected voxel points is determined as a connected component, and the local minimum point found is marked as the birthpoint of this connected component, although this is not limited here.

[0095] S2072: When two connected components are connected and merged for the first time at the current grayscale value, the connected component with the higher birth value is terminated, and the current grayscale value is recorded as the extinction point corresponding to the connected component with the higher birth value, where the birth value refers to the grayscale value corresponding to the birth point of the connected component;

[0096] Among them, the higher birth value refers to the comparison of the grayscale values ​​corresponding to the birth points of the two connected components to be merged. The grayscale value with the larger value is the higher birth value. This birth value reflects a characteristic state when the connected component is initially formed.

[0097] Specifically, this step is performed after the birth points of each connected component have been determined and the changes in connected components over the evolution of the topological structure are observed. As the nested simplex-complex sequence is further analyzed based on grayscale values ​​(for example, as the grayscale value of the voxel point continues to rise, the topological structure continues to change), when two previously independent connected components begin to connect with each other and merge into a larger connected area, this important topological change information needs to be recorded. At this time, the grayscale values ​​corresponding to the birth points of the two connected components are compared, and the connected component with the higher birth value is found. The current grayscale value at the time of merging is marked as the extinction point of this connected component. This clarifies the existence interval of the connected component throughout the evolution of the topological structure, that is, from the grayscale value corresponding to the birth point to the grayscale value corresponding to the extinction point. This facilitates subsequent operations such as constructing a persistence graph to more comprehensively analyze the stability and change characteristics of the topological structure.

[0098] In some specific embodiments, the status of each connected component and the connectivity between them are monitored in real time. When a connection and merger between two connected components is detected, the grayscale value data corresponding to the birth points of each of the two connected components is obtained. By comparing the grayscale values, the connected component with the higher birth value is determined. The grayscale value corresponding to the voxel at that moment, i.e., the current grayscale value, is then recorded and marked as the extinction point of the connected component with the higher birth value. A dedicated data structure can be established to store information such as the birth and extinction points of connected components for subsequent query and use, which is not limited here.

[0099] S2073. Construct a persistence graph based on the birth point and the corresponding extinction point, wherein the abscissa of the persistence graph is the grayscale value of the birth point, and the ordinate of the persistence graph is the grayscale difference between the birth point and the corresponding extinction point.

[0100] Specifically, after the birth point and corresponding extinction point of each connected component have been determined, the grayscale value of the birth point of each connected component is used as the value of the horizontal coordinate, and the grayscale difference between its birth point and the corresponding extinction point is calculated as the value of the vertical coordinate. A point is determined in the plane rectangular coordinate system to represent this connected component, and all the points corresponding to the connected components are drawn in the same coordinate system to form a persistence graph.

[0101] As can be seen, during the analysis of connected components, the birth points of connected components are determined by local minima, connected components are formed based on voxel connectivity, and when connections merge, the extinction points are determined based on the birth values, and a persistence graph is constructed. This method, based on recording key nodes of topological evolution, deeply transforms image grayscale information into a quantitative description of topological structural features. The persistence graph intuitively displays the existence range and stability of connected components. By utilizing information related to the birth and extinction points, it can effectively evaluate the topological changes of the particle structure under different grayscale value changes, making the selection of seed points more accurate.

[0102] Continuing with the previous example, when performing persistence calculations, although the grayscale value of a single voxel is fixed, the focus is on the evolution of the topological structure at different grayscale levels (determined by the ascending order of the voxels). For example, the topological structure within a local region may begin to form at a low grayscale value. As the grayscale value increases (that is, as voxels with higher grayscale values ​​are processed), this topological structure may merge, disappear, or undergo other changes. Persistence calculations quantify this change in topological structure as the grayscale values ​​of voxels are processed in ascending order. Points that are relatively stable locally (i.e., the topological structure corresponding to the local minimum) and meet certain stability requirements (persistence greater than a threshold) are used as seed points. The starting position (local minimum) of the topological structure of the local region where these seed points are located is of great significance throughout the grayscale evolution from low to high, making them suitable as a starting reference for particle identification.

[0103] The following describes an exemplary rock specimen-based particle structure map generation system 400 provided in an embodiment of the present application. Figure 4 Schematic diagram of an exemplary hardware structure of a rock specimen-based particle structure atlas generation system 400 provided in an embodiment of the present application.

[0104] In some embodiments, the rock specimen-based particle structure atlas generation system 400 is a computer device or the rock specimen-based particle structure atlas generation system 400 includes a computer device. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface, and in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, the method in the embodiment of the present application is implemented.

[0105] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0106] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0107] As used in the above embodiments, the term when may be interpreted to mean if or after or in response to determining or in response to detecting, depending on the context. Similarly, the phrases upon determining or upon detecting (the stated condition or event) may be interpreted to mean if determining or in response to determining or upon detecting (the stated condition or event) or in response to detecting (the stated condition or event), depending on the context.

[0108] In the above embodiments, all or part of the embodiments can be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of this application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that integrates one or more available media. The available medium can be magnetic media (e.g., floppy disk, hard disk, tape), optical media (e.g., DVD), or semiconductor media (e.g., solid-state drive).

[0109] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for generating a particle structure map based on rock specimens, characterized in that: include: Performing three-dimensional scanning on the rock specimen to obtain a sequence of cross-sectional images of the internal structure of the rock specimen; Identifying a seed point from the cross-sectional image sequence based on image grayscale values, and setting the seed point as an initial voxel; wherein: the grayscale values ​​in the cross-sectional image sequence are used as height values ​​to construct a three-dimensional height field, wherein the position of each voxel point is determined by the corresponding spatial coordinates, and the height value of the voxel point is determined by the corresponding grayscale value; sorting the voxel points in the three-dimensional height field in ascending order; In ascending order, voxel points are added as vertices to the simplex complex, and each voxel point is constructed as a zero-dimensional simplex structure; Searching for adjacent voxel points of a voxel point within a preset neighborhood of the three-dimensional height field, and connecting two voxel points whose distance is less than a distance threshold to construct a one-dimensional simplex; When the endpoints of the one-dimensional simplex edges are sequentially connected end to end to form a closed loop, the region enclosed by the closed loop is constructed as a two-dimensional simplex surface, and the two-dimensional simplex surface is added to the simplex complex; forming a nested simplex complex sequence, wherein each layer of the simplex complex sequence represents a topological structure whose grayscale value is less than or equal to the grayscale value of the current voxel point; Performing persistence calculation on the nested simplex-complex sequence, and calculating a local minimum point whose persistence is greater than the persistence threshold as a seed point; Comparing the grayscale values ​​of adjacent voxels of the initial voxel, and classifying the adjacent voxels that meet a preset growth threshold condition as the same particle; Performing feature analysis on all acquired particles to obtain digital features, including: morphological characteristic parameters of the particles, spatial distribution characteristics of the particles, contact relationship network between particles, and gap network between particles; Projecting the particles onto a two-dimensional plane according to the digital features to obtain a projection image; An analysis report and a visualization map are generated, wherein the analysis report includes morphological characteristic parameters of the particles, spatial distribution characteristics of the particles, contact relationship networks between particles, and gap networks between particles, and the visualization map includes the projection image.

2. The method according to claim 1, characterized in that The step of performing persistence calculation on the nested simplex-complex sequence specifically includes: A new connected component is generated at a local minimum point, which is recorded as the birth point of the connected component; the connected component is formed by interconnected voxel points within the nested simplex-complex sequence, and the local minimum point is the point with the smallest grayscale value in the pre-divided local area; When two connected components are connected and merged for the first time at the current grayscale value, the connected component with the higher birth value is terminated, and the current grayscale value is recorded as the extinction point corresponding to the connected component with the higher birth value, and the birth value refers to the grayscale value corresponding to the birth point of the connected component; A persistence graph is constructed based on the birth point and the corresponding extinction point, wherein the abscissa of the persistence graph is the grayscale value of the birth point, and the ordinate of the persistence graph is the grayscale difference between the birth point and the corresponding extinction point.

3. The method according to claim 1, characterized in that The step of performing feature analysis on all acquired particles to obtain digital features specifically includes: Calculating the ratio of the surface area to the volume of each of the particles to obtain a sphericity parameter; The ratio of the longest axis to the shortest axis of each particle is calculated to obtain the elongation parameter.

4. The method according to claim 1, wherein The step of performing feature analysis on all acquired particles to obtain digital features specifically includes: Counting the particle size distribution of the particles; According to the coordinate information of the particles in the three-dimensional space, a position density distribution map of the particles in the three-dimensional space is constructed to obtain the spatial distribution characteristics of the particles.

5. The method according to claim 1, wherein The step of performing feature analysis on all acquired particles to obtain digital features specifically includes: Identify contact areas between adjacent particles; Calculating the area and perimeter of the contact area; classifying the contact region as point contact, line contact, or surface contact according to the area and the perimeter; Based on the contact relationships between the particles, a contact relationship network between the particles is constructed.

6. The method according to claim 1, characterized in that The step of performing feature analysis on all acquired particles to obtain digital features specifically includes: identifying interstitial regions between the particles; Counting the size distribution of the void area; A network of voids between the particles is constructed.

7. A particle structure atlas generation system based on rock specimens, characterized in that: The rock specimen-based particle structure map generation system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the rock specimen-based particle structure map generation system to execute the method described in any one of claims 1 to 6.

8. A computer program product comprising instructions, characterized in that When the computer program product is run on a system for generating a particle structure map based on a rock specimen, the system is enabled to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a system for generating a particle structure map based on a rock specimen, the system is caused to execute the method according to any one of claims 1 to 6.

Images