A full-field digital thin-section analysis system and method for geological core analysis

Abstract

The present invention relates to a full-view digital thin-section analysis system and method for geological core analysis, belonging to the technical field of geological exploration. The system includes: a matrix construction module for obtaining multi-band images of core thin sections and establishing a multi-scale image acquisition matrix; a first function construction module for constructing a three-dimensional spatial structure function; a feature fusion module for generating a feature fusion matrix; a second function construction module for defining a stitching optimization function and calculating the stitching optimization result; an integrity evaluation module for evaluating the continuity of the three-dimensional structure of core thin sections; an image update module for dynamically adjusting the learning rate and the time decay coefficient based on the gradient direction of the integrity evaluation function, and updating the image parameters until convergence; and a three-dimensional image generation module for selecting key feature points of core thin sections and generating a three-dimensional visualization image of core thin sections through a spatial distance attenuation coefficient. This system can improve the accuracy of digital thin-section analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of geological exploration, and particularly to a full-field digital thin-section analysis system, method, electronic device, and non-transitory computer-readable storage medium for geological core analysis. Background Art

[0002] Currently, as a key means in geological exploration, reservoir evaluation, and oil and gas reservoir description, core thin-section analysis widely uses the traditional method of combining optical microscopy with manual identification, supplemented by digital image stitching and image recognition algorithms for image acquisition and analysis.

[0003] However, existing digital thin-section analysis methods generally have problems such as strong field-of-view limitations, lack of spatial structure information, and fragmented processing flows. Most methods are still based on local field-of-view stitching, making it difficult to restore the continuous spatial structure and phase distribution characteristics of core thin sections at the macroscopic scale, resulting in distortion or loss of core structure information during the stitching process. Summary of the Invention

[0004] The present invention aims at the technical problems existing in the prior art and provides a full-field digital thin-section analysis system, method, electronic device, and non-transitory computer-readable storage medium for geological core analysis that can improve the accuracy of digital thin-section analysis.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] The present invention provides a full-field digital thin-section analysis system for geological core analysis, the system comprising:

[0007] A matrix construction module, configured to obtain multi-band images of a core thin section and establish a multi-scale image acquisition matrix;

[0008] A first function construction module, configured to suppress noise and enhance edge features using a gradient operator, and construct a three-dimensional spatial structure function in combination with a feature point distribution function;

[0009] A feature fusion module, configured to perform a joint transformation on the multi-scale image acquisition matrix and the three-dimensional spatial structure function, calculate a fusion weight through feature similarity measurement, and generate a feature fusion matrix;

[0010] A second function construction module, configured to define a stitching optimization function and calculate a stitching optimization result according to reference point coordinates and a smoothing coefficient;

[0011] An integrity evaluation module, configured to calculate the absolute value of the gradient of the stitching optimization result, and evaluate the continuity of the three-dimensional structure of the core thin section in combination with a structure integrity measurement function;

[0012] An image update module, configured to dynamically adjust the learning rate and the time decay coefficient based on the gradient direction of the integrity evaluation function, and update the image parameters until convergence;

[0013] A three-dimensional image generation module, configured to select key feature points of the core thin section, and generate a three-dimensional visualization image of the core thin section through a spatial distance decay coefficient.

[0014] Optionally, the matrix construction module is further configured to:

[0015] Obtain the intensity distribution of the multi-band image;

[0016] Determine each image matrix element according to the intensity distribution of the multi-band image;

[0017] Construct the multi-scale image acquisition matrix according to each of the image matrix elements.

[0018] Optionally, the multi-scale image acquisition matrix is expressed as:

[0019] ;

[0020] Where M is the multi-scale image acquisition matrix, is the image matrix element, is the weight coefficient under different image channels or scales, is the image intensity at the corresponding position and band, is the wavelength parameter, is the dimension of the image matrix.

[0021] Optionally, the first function construction module is further configured to:

[0022] Perform a three-dimensional integration on the multi-scale image acquisition matrix, and perform noise suppression through a gradient operator to obtain a corresponding integral term;

[0023] Superimpose a feature point distribution function on the integral term to obtain a spatial construction function.

[0024] Optionally, the spatial construction function is expressed as:

[0025] ;

[0026] Where is the spatial structure reconstruction function, is the voxel representation of the multi-scale image acquisition matrix, is the gradient operator, is the feature point distribution function, m is the number of feature points, are the first adjustment parameter, the second adjustment parameter, and the third adjustment parameter respectively.

[0027] Optionally, the feature fusion module is further configured to:

[0028] Obtain a spatial structure function for describing the structural integrity and three-dimensional distribution of the core thin section;

[0029] Define a joint transformation operator for spatially aligning the multi-scale image acquisition matrix and the spatial structure function;

[0030] Calculate an attenuation factor based on the feature distance, and generate the fused image feature values;

[0031] Generate the feature fusion matrix according to the image feature values;

[0032] Optionally, the second function construction module is further configured to:

[0033] Calculate the corresponding spatial attenuation factor according to the reference point coordinates;

[0034] Determine the stitching optimization function according to the spatial attenuation factor, the smoothing coefficient, and the feature fusion matrix;

[0035] Optionally, the integrity evaluation module is further configured to:

[0036] Obtain the local pore volume, the connected path length, and the fracture density of the core thin section;

[0037] Determine the structural integrity measurement function according to the local pore volume, the connected path length, and the fracture density of the core thin section;

[0038] Optionally, the three-dimensional image generation module is further configured to:

[0039] Determine the corresponding distance attenuation factor according to the key feature points of the core thin section and the distance attenuation coefficient;

[0040] Determine the three-dimensional visualization image of the core thin section according to the distance attenuation factor and the stitching optimization function value calculated by the stitching optimization function;

[0041] The present invention also provides a geological core analysis full-field digital thin section analysis method, and the method includes:

[0042] Obtain multi-band images of the core thin section, and establish a multi-scale image acquisition matrix;

[0043] Use a gradient operator to suppress noise and enhance edge features, and combine with a feature point distribution function to construct a three-dimensional spatial structure function;

[0044] Perform a joint transformation on the multi-scale image acquisition matrix and the three-dimensional spatial structure function, calculate the fusion weight through feature similarity measurement, and generate a feature fusion matrix;

[0045] Define the splicing optimization function and calculate the splicing optimization result according to the reference point coordinates and the smoothing coefficient;

[0046] Calculate the absolute value of the gradient of the splicing optimization result, and combine the structural integrity measurement function to evaluate the continuity of the three-dimensional structure of the core thin section;

[0047] Based on the gradient direction of the integrity evaluation function, dynamically adjust the learning rate and the time decay coefficient, and update the image parameters until convergence;

[0048] Select the key feature points of the core thin section, and generate the three-dimensional visualization image of the core thin section through the spatial distance decay coefficient.

[0049] In addition, to achieve the above object, the present invention also provides an electronic device, including: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby implementing a method for analyzing a full-field digital thin section of geological core as described above.

[0050] In addition, to achieve the above object, the present invention also provides a non-transitory computer-readable storage medium, in which a computer software program is stored, and when the computer software program is executed by a processor, it implements a method for analyzing a full-field digital thin section of geological core as described above.

[0051] The beneficial effects of the present invention are:

[0052] (1) By establishing a multi-scale image acquisition matrix, the present invention effectively integrates image information of different scales and bands within the local field of view, forms a continuous and complete core image set, provides high-quality raw data support for subsequent structure analysis and three-dimensional modeling, and significantly improves the spatial coverage of the thin section image.

[0053] (2) Through the spatial structure reconstruction function, the present invention not only considers the overall image information (through the integral term), but also introduces the explicit feature point distribution, and effectively reconstructs the microscopic structure morphology through the gradient regulation mechanism, overcomes the limitations of traditional methods that only rely on image segmentation or texture judgment, and improves the accuracy and robustness of three-dimensional reconstruction.

[0054] (3) By constructing a feature fusion matrix, the present invention jointly encodes the image features and the structure information, and introduces a feature distance and weight regulation mechanism, further enhancing the contrast of key structures, mineral boundaries, and fracture features, which is helpful to achieve more accurate structure recognition and classification analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a scene diagram of a method for analyzing a full-field digital thin section of geological core provided by the present invention;

[0056] Figure 2 Schematic diagram of the structure of a full - view digital thin - section analysis system for geological core analysis provided by the present invention;

[0057] Figure 3 Flowchart of a full - view digital thin - section analysis method for geological core analysis provided by the present invention;

[0058] Figure 4 Schematic diagram of the hardware structure of a possible electronic device provided by the present invention;

[0059] Figure 5 Schematic diagram of the hardware structure of a possible computer - readable storage medium provided by the present invention. Detailed implementation manners

[0060] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0061] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0062] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or having more advantages than other embodiments. In order to enable any person skilled in the art to implement and use the present invention, the following description is given. In the following description, details are set forth for the purpose of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other instances, well - known structures and processes are not described in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope that conforms to the principles and features disclosed in the present invention.

[0063] Please refer to Figure 1 , Figure 1 Scene diagram of a full - view digital thin - section analysis method for geological core analysis provided by the present invention. As Figure 1As shown, the terminal is connected to the server through a network, for example, through a wired or wireless network connection, etc. Among them, the terminal may include, but is not limited to, portable terminals such as mobile phones and tablets installed with various network platform applications, as well as fixed terminals such as computers, query machines, and advertising machines. Among them, the server provides various business services for users, including service push servers, user recommendation servers, etc.

[0064] It should be noted that Figure 1 The scenario diagram of a full-field digital thin-section analysis method for geological core analysis shown is only an example. The terminal, server, and application scenarios described in the embodiments of the present invention are for more clearly explaining the technical solutions of the embodiments of the present invention, and do not limit the technical solutions provided by the embodiments of the present invention. Those of ordinary skill in the art know that with the evolution of the system and the emergence of new business scenarios, the technical solutions provided by the embodiments of the present invention are equally applicable to similar technical problems.

[0065] Among them, the terminal can be used for:

[0066] Obtain multi-band images of core thin sections and establish a multi-scale image acquisition matrix;

[0067] Use the gradient operator to suppress noise and enhance edge features, and combine with the feature point distribution function to construct a three-dimensional spatial structure function;

[0068] Perform a joint transformation on the multi-scale image acquisition matrix and the three-dimensional spatial structure function, calculate the fusion weight through feature similarity measurement, and generate a feature fusion matrix;

[0069] Define a stitching optimization function and calculate the stitching optimization result according to the reference point coordinates and the smoothing coefficient;

[0070] Calculate the absolute value of the gradient of the stitching optimization result, and combine with the structure integrity measurement function to evaluate the continuity of the three-dimensional structure of the core thin section;

[0071] Based on the gradient direction of the integrity evaluation function, dynamically adjust the learning rate and the time decay coefficient, and update the image parameters until convergence;

[0072] Select the key feature points of the core thin section, and generate a three-dimensional visualization image of the core thin section through the spatial distance attenuation coefficient.

[0073] Please refer to Figure 2 , Figure 2 , which is a schematic structural diagram of a full-field digital thin-section analysis system for geological core analysis provided by the present invention.

[0074] As Figure 2 shown, a full-field digital thin-section analysis system for geological core analysis proposed in the embodiments of the present invention includes:

[0075] The matrix construction module 201 is configured to obtain multi - band images of core thin sections and establish a multi - scale image acquisition matrix;

[0076] The first function construction module 202 is configured to suppress noise and enhance edge features using a gradient operator, and construct a three - dimensional spatial structure function in combination with a feature point distribution function;

[0077] The feature fusion module 203 is configured to perform a joint transformation on the multi - scale image acquisition matrix and the three - dimensional spatial structure function, calculate fusion weights through feature similarity measurement, and generate a feature fusion matrix;

[0078] The second function construction module 204 is configured to define a stitching optimization function and calculate a stitching optimization result according to reference point coordinates and a smoothing coefficient;

[0079] The integrity evaluation module 205 is configured to calculate the absolute value of the gradient of the stitching optimization result, and evaluate the continuity of the three - dimensional structure of the core thin section in combination with a structural integrity measurement function;

[0080] The image update module 206 is configured to dynamically adjust the learning rate and the time decay coefficient based on the gradient direction of the integrity evaluation function, and update the image parameters until convergence;

[0081] The three - dimensional image generation module 207 is configured to select key feature points of the core thin section and generate a three - dimensional visualization image of the core thin section through a spatial distance decay coefficient.

[0082] In some embodiments, the matrix construction module 201 may further be configured to:

[0083] Obtain the intensity distribution of the multi - band image;

[0084] Determine each image matrix element according to the intensity distribution of the multi - band image;

[0085] Construct the multi - scale image acquisition matrix according to each image matrix element.

[0086] In some embodiments, the multi - scale image acquisition matrix can be expressed as:

[0087] ;

[0088] where M is the multi - scale image acquisition matrix, is the image matrix element, is the weight coefficient under different image channels or scales, is the image intensity at the corresponding position and band, is the wavelength parameter, is the dimension of the image matrix.

[0089] In a specific implementation, M is the original data container for subsequent spatial reconstruction and feature fusion. Each element represents the fusion result of the image intensity at a specific position. It is a two-dimensional matrix of n×n, and each element corresponds to an image pixel. is used to adjust the importance of images at different scales / bands, which may be set according to image clarity, contrast, or preset rules. represents the image brightness or reflectivity at the coordinate (x, y) and wavelength under the conditions of the k-th and l-th channels or scales. is usually used to represent a certain imaging band, such as visible light, ultraviolet, infrared, X-ray, etc.; it is assumed to be a fixed band here . n represents the resolution of the image in the horizontal and vertical directions (i.e., the number of pixels), and can also be understood as the side length of the matrix.

[0090] represents the original image data at multiple scales or channels (such as different magnifications, different imaging instruments, different spectra), and each data source is fused with different weights added. The image information from all these different sources is weighted and superimposed at each pixel point (x, y) to form an image matrix M with a unified expression, providing a unified input for subsequent spatial reconstruction and feature analysis. By controlling and selecting different , this model can be adapted to different types of core image data acquisition devices and modes (such as microscope images, multispectral scanning images, etc.).

[0091] As can be seen from the above, the process of this formula in the present invention is equivalent to "fusion calibration" for different image sources, and the purposes are: to unify the image scale and spectral differences, eliminate the imaging differences between devices; to improve the expression ability of the local features of the image and retain the key information in each scale; to provide a high-information input matrix M for subsequent three-dimensional structure reconstruction.

[0092] In some embodiments, the first function construction module 202 is further configured to:

[0093] perform a three-dimensional integration on the multi-scale image acquisition matrix and suppress noise through a gradient operator to obtain a corresponding integral term;

[0094] superimpose a feature point distribution function on the integral term to obtain a spatial construction function.

[0095] In some embodiments, the spatial construction function can be expressed as:

[0096] ;

[0097] where, is the spatial structure reconstruction function, is the voxel representation of the multi-scale image acquisition matrix, is the gradient operator, is the feature point distribution function, m is the number of feature points, They are the first adjustment parameter, the second adjustment parameter and the third adjustment parameter respectively.

[0098] In the specific implementation, Describes the structural expression value of a point in three-dimensional space, is the expanded form of the image matrix in three-dimensional space, Indicates the local spatial change rate of the image (such as edge, texture, and structural mutation). Emphasizes areas of dramatic change, often indicating texture complexity or boundary clarity. Characterizes the distribution of key structures (such as cracks, mineral boundaries, etc.) in three-dimensional space, m is the total number of feature points, Control the influence of the integral term in the structure construction (dominate the overall situation), Control the sensitivity to image gradients (such as edge suppression or enhancement), Control the influence of feature point items on the structure (dominant local).

[0099] Image volume integration based on gradient modulation , the purpose is to extract a smooth, continuous structure distribution from the original image. This term extracts structural features from the image intensity matrix M, using an exponential decay term Suppress high gradient areas (such as noise or sharp edges) and enhance areas with strong continuity and clear structure. Acts as an overall strength factor to modulate the influence of global background information.

[0100] The distribution of feature points detected manually or automatically is explicitly added to ensure that important features are not weakened by the smoothing process during the structure reconstruction process and to enhance the expression of key details. It is the representation of feature points in three-dimensional space (such as Gaussian kernel function, Dirac function, etc.). Control the contribution of these key feature points to the overall structure.

[0101] This structure function is a "global-local fusion model": the global structure is constructed by M+ gradient control terms, emphasizing continuity; the local features are constructed by Implantation, retaining key point information of geological structure; three adjustment parameters It can be dynamically adjusted through experiments or optimization to adapt to the characteristics of different core image data.

[0102] In the above manner, the present invention can be used for tasks such as three-dimensional core reconstruction, structural analysis, and lithofacies identification, providing a spatial data basis for subsequent feature fusion and structural integrity assessment; decoupled from the image preprocessing step, it has good modularity and controllability.

[0103] In some embodiments, the feature fusion module 203 is further configured to:

[0104] Obtain a spatial structure function for describing the structural integrity and three-dimensional distribution of the core thin section;

[0105] Define a joint transformation operator for spatially aligning the multi-scale image acquisition matrix and the spatial structure function;

[0106] Calculate an attenuation factor based on the feature distance and generate the fused image feature values for each;

[0107] Generate the feature fusion matrix according to the image feature values for each.

[0108] In some embodiments, the feature fusion matrix can be expressed as:

[0109] ;

[0110] where F is the feature fusion matrix, is the fused image feature value, is the fusion weight coefficient, is the transformation operator acting on the image and the structure, is the feature distance (similarity metric), is the distance attenuation coefficient.

[0111] In specific implementation, each element is the fused composite image feature, forming an intermediate expression layer for subsequent stitching and modeling. is used to adjust the contribution degree of the k-th and l-th feature fusion units to the whole, and may be set according to the image region saliency or reliability. is a function for simultaneously performing feature extraction, transformation, or encoding processing on the image information M and the structure information S, such as PCA, CNN, filter bank, attention mechanism, etc. Measures the feature difference between the fusion unit and the central region, and is used for spatial weight suppression to avoid the influence of noise. Controls the influence degree of the feature distance on the fusion contribution, and the larger the value, the more sensitive the distance. Represents the spatial resolution of the fusion matrix, usually smaller than the original image size, equivalent to a dimensionality reduction or high-level expression layer.

[0112] This formula uses the transformation operator , integrate and analyze the information in the image intensity matrix M and the spatial structure function S; this operation can support the fusion at different scales and different semantic levels. For example, M captures the image texture and brightness differences, S emphasizes the structural integrity and three-dimensional distribution, and T realizes joint dimensionality reduction, feature reconstruction, or significant region extraction. Control the contribution of regions that are far away or have large differences to the current feature points, reduce the interference from low-correlation regions, and enhance local consistency. It can be automatically set according to image noise, structural complexity, or specific algorithms (such as attention distribution); thus, the fusion intensity of different regions is controllable.

[0113] This feature fusion matrix simultaneously utilizes the image brightness / texture information M and the spatial structure information S to achieve the integration of multi-source information, and uses and Control the fusion radius and similarity weights to achieve local sensitive modulation, and the output dimension is usually more abstract than the original image as a high-level representation, realizing the conversion of feature dimensions. Provide a refined feature basis for the stitching optimization function O(x,y), and it has strong adaptability subsequently.

[0114] In summary, the feature fusion matrix of the present invention is a "comprehensive expression" of images and structures, and can be used for image stitching and matching, the input feature layer of model training, spatial pattern recognition (such as fracture network and bedding structure extraction), and improving feature robustness and spatial consistency.

[0115] In some embodiments, the second function construction module 204 is further configured to:

[0116] Calculate the corresponding spatial attenuation factor according to the reference point coordinates;

[0117] Determine the stitching optimization function according to the spatial attenuation factor, the smoothing coefficient, and the feature fusion matrix.

[0118] In some embodiments, the stitching optimization function can be expressed as:

[0119] ;

[0120] where is the stitching optimization function, is the value in the feature fusion matrix, is the stitching coefficient, is the spatial attenuation parameter, and are the reference point coordinates.

[0121] In specific implementation, represents the pixel value or feature value after fusion output at the position (x,y), and is used to reconstruct the complete view. Contains image and structural information. Controls the contribution intensity from the reference point which is related to the feature confidence or position error. , is the position of the key stitching point or feature point in the original image, defining the central region of the overlap / stitching. Controls , the influence of the distance from the reference point in space on the contribution degree, determining the "blur range" of the fusion.

[0122] Each reference point serves as a "local fusion center"; suppresses the interference of regions far from the center through a spatial Gaussian-type weight ; the influence ranges of each reference point may overlap, thus achieving smooth transition and continuous stitching. The stitching coefficient can be set according to the following information: local image quality (blur, exposure), image overlap degree, importance of structural features, supports non-uniform stitching strategies, and improves edge consistency and geometric alignment.

[0123] In the present invention, this stitching function has multiple values in the analysis of geological core images: eliminating stitching gaps: through spatial weighting, buffering the discontinuity between different image blocks; enhancing structural consistency: using F(x,y) to retain the spatial structural features, making the stitched image not only visually coherent but also maintaining consistent structural information; adapting to complex core data: supporting the unified fusion of images under different resolutions, sampling regions, and imaging conditions.

[0124] In some embodiments, the integrity assessment module 205 is further configured to:

[0125] Obtain the local pore volume, connected path length, and fracture density of the core slice;

[0126] Determine the structural integrity metric function according to the local pore volume, connected path length, and fracture density of the core slice.

[0127] In some embodiments, the structural integrity metric function can be expressed as:

[0128] ;

[0129] where Q is the structural integrity metric function, is the three-dimensional gradient, is the evaluation weight parameter, is the structural integrity metric function.

[0130] In a specific implementation, the gradient of the image change → Represents the structural boundary and texture clarity; the integrity metric R(x, y, z) of the spatial structure → Measures whether the local structure is complete and continuous; the attenuation control parameter η → Controls the impact of the structure missing area on the overall score. The finally obtained Q is an overall evaluation index, which can be used for quality control, stitching effect evaluation, or the regularization term in the optimization objective function.

[0131] Q is used to measure the integrity and quality of the overall spatial structure or image. Represents the change rate of the image in the three-dimensional space after stitching optimization, reflecting the edge / texture clarity. Represents the degree of integrity loss at the spatial point (x, y, z), such as holes, blurs, outliers, etc. Controls the impact of the structure missing area on the overall score. The larger the value, the stronger the suppression of the missing area.

[0132] The gradient is used as an index of structural strength. The larger it is → Indicates that the image structure changes strongly, with obvious edges and clear textures; suitable for identifying bedding, fractures, striations, etc. in core images; if the gradient is small, it indicates that the area is smooth, which may be a structure missing or fusion error area. R(x, y, z) suppresses the interference of the missing area. When R is larger → Indicates that the regional structure is incomplete (such as missing, misaligned, ghosting). Approaches zero, thereby reducing the contribution of this area to the overall evaluation; effectively avoiding the "misleading" of the overall evaluation by the defective area. The three-dimensional integral ensures global analysis, through triple integral Statistics the integrity performance within the entire volume or image domain; applicable to multi-slice data, three-dimensional structure modeling, etc.

[0133] In summary, the present invention provides a quantifiable index for integrating image quality. Areas with low Q may have stitching errors or structure loss, which can be used for iterative optimization in subsequent steps. As part of the loss function, the results generated by different stitching algorithms or parameter combinations can be objectively compared through Q.

[0134] In some embodiments, the state of the (t + 1)-th iteration of the image parameters can be expressed as:

[0135] ;

[0136] Wherein, is the state of the t-th iteration, is the state of the (t + 1)-th iteration, is the learning rate or update step size, is the time decay coefficient, is the gradient of the integrity function Q, is the time decay factor.

[0137] In a specific implementation, this formula defines an iterative strategy for state optimization based on the integrity evaluation function Q. Its goal is to iteratively update and optimize in a directed manner based on the image / structure integrity evaluation results while preserving the existing state. The following are introduced: gradient guidance ( ): indicating how to adjust the state to improve the overall integrity; time decay factor : controlling the iterative amplitude to decrease over time and tend to converge; step size parameter (δ): adjusting the intensity of each update step.

[0138] Gradient-driven update , if the integrity function Q is regarded as the objective function, this formula is a typical gradient ascent method. The gradient indicates in which directions the current state can improve the structural integrity; actually, can be the partial derivative with respect to spatial position, image pixel value, or structural parameter.

[0139] Learning rate control , determining the adjustment amplitude of the state for each step; often needs to be set to a small value to ensure the stable convergence of the system; an adaptive adjustment mechanism can be adopted to dynamically adjust δ according to the gradient magnitude.

[0140] Time decay mechanism , the update speed is relatively fast in the initial stage, which is beneficial to quickly jump out of local minima; as the number of iterations t increases, the decay term decreases, making the optimization tend to be stable; it is equivalent to a simulated annealing-like strategy, avoiding oscillatory convergence and enhancing robustness.

[0141] The initial state I(0) can be the initial image, structural reconstruction sketch, or blank model. The integrity Q is evaluated and analyzed using the integrity evaluation function, and the gradient is calculated to obtain the sensitive directions of the current state for improving Q, and the state I(t + 1) is updated to update the model or image using a well-controlled step size and decay coefficient.

[0142] In summary, the present invention makes a directional adjustment for the integrity objective, avoiding blind updates. The time decay strategy ensures fast speed in the initial stage and stability in the later stage, and can be embedded in neural network training, image enhancement, 3D reconstruction, etc. It is applicable to various state update processes, including image data, transformation matrices, structural models, etc.

[0143] In some embodiments, the 3D image generation module 207 is further configured to:

[0144] Determine the corresponding distance decay factor according to the key feature points of the core thin section and the distance decay coefficient;

[0145] Determine the 3D visualization image of the core thin section according to the distance decay factor and the stitching optimization function value calculated by the stitching optimization function.

[0146] In some embodiments, the three-dimensional visualization image of the core thin section can be represented as:

[0147] ;

[0148] Where is the intensity or feature map of the three-dimensional visualization image, is the value of the stitching optimization function, is the weight of the i-th feature point, is the spatial distance attenuation coefficient, is the spatial position of the feature point, is the distance metric.

[0149] In a specific implementation, this function is used to generate a three-dimensional visualization image or a feature map result based on the three-dimensional stitching optimization result O(x, y, z). It fuses the influences of multiple feature points into the three-dimensional space, such that: the visualization effect focuses on the key structural regions; different feature points control their influence ranges with weights and the spatial distance attenuation σ; a weight-guided and space-sensitive three-dimensional image mapping model is constructed.

[0150] Based on the spatial weighting mechanism of the feature points, each feature point will "radiate" influence within a certain range around it; the closer the distance, the greater the influence on the current point (x, y, z), and vice versa; the attenuation function controls the attenuation of the influence intensity with the spatial distance; similar to the Gaussian weight kernel, but without the square term, so the influence radius is larger and the transition is gentler.

[0151] Feature point-guided focus enhancement, where feature points are often the key analysis regions (such as fracture zones, pore clusters, sedimentary structures, etc.); by controlling its "visualization weight", enhanced display of key regions and weakened processing of non-key regions are achieved; it is beneficial for subsequent expert identification, model training, or interactive analysis.

[0152] Combining the original value with the weighted kernel mapping, this mapping does not "replace" the original stitching data, but "modulates" it and presents it; making the result have both the authenticity of the original structure and the prominence of the feature regions.

[0153] In summary, the present invention can enhance the display effect of the regions around key feature points, help geological experts quickly locate important regions in the three-dimensional structure, construct a heat field and a region of interest (ROI) mask, which can be output as a feature map for use in machine learning and neural networks, and can stack V(x, y, z) of multiple channels to construct a multi-view visualization model.

[0154] Please refer to Figure 3, a flowchart of a full - view digital thin - section analysis method for geological core analysis according to the present invention is provided, including the following steps:

[0155] Step 301, obtain multi - band images of core thin sections and establish a multi - scale image acquisition matrix;

[0156] Step 302, use a gradient operator to suppress noise and enhance edge features, and combine with a feature point distribution function to construct a three - dimensional spatial structure function;

[0157] Step 303, perform a joint transformation on the multi - scale image acquisition matrix and the three - dimensional spatial structure function, calculate the fusion weight through feature similarity measurement, and generate a feature fusion matrix;

[0158] Step 304, define a stitching optimization function and calculate the stitching optimization result according to the reference point coordinates and the smoothing coefficient;

[0159] Step 305, calculate the absolute value of the gradient of the stitching optimization result, and combine with a structure integrity measurement function to evaluate the continuity of the three - dimensional structure of the core thin section;

[0160] Step 306, dynamically adjust the learning rate and the time decay coefficient based on the gradient direction of the integrity evaluation function, and update the image parameters until convergence;

[0161] Step 307, select key feature points of the core thin section, and generate a three - dimensional visualization image of the core thin section through a spatial distance attenuation coefficient.

[0162] It should be noted that for the specific embodiments and beneficial effects of the above steps 301 - 307, please refer to the description content of parts 201 - 207 of the module, which will not be elaborated here.

[0163] Please refer to Figure 4 , Figure 4 , which is a schematic diagram of an embodiment of an electronic device provided by an embodiment of the present invention. As Figure 4 shown, an embodiment of the present invention provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and operable on the processor 420. When the processor 420 executes the computer program 411, the following steps are implemented:

[0164] Obtain multi - band images of core thin sections and establish a multi - scale image acquisition matrix;

[0165] Use a gradient operator to suppress noise and enhance edge features, and combine with a feature point distribution function to construct a three - dimensional spatial structure function;

[0166] Perform a joint transformation on the multi-scale image acquisition matrix and the three-dimensional spatial structure function, calculate the fusion weights through feature similarity measurement, and generate a feature fusion matrix;

[0167] Define a stitching optimization function and calculate the stitching optimization result according to the reference point coordinates and the smoothing coefficient;

[0168] Calculate the absolute value of the gradient of the stitching optimization result, and combine it with the structure integrity measurement function to evaluate the continuity of the three-dimensional structure of the core thin section;

[0169] Based on the gradient direction of the integrity evaluation function, dynamically adjust the learning rate and the time decay coefficient, and update the image parameters until convergence;

[0170] Select the key feature points of the core thin section, and generate a three-dimensional visualization image of the core thin section through the spatial distance decay coefficient.

[0171] Please refer to Figure 5 , Figure 5 which is a schematic diagram of an embodiment of a computer-readable storage medium provided by an embodiment of the present invention. As Figure 5 shown, this embodiment provides a computer-readable storage medium 500, on which a computer program 411 is stored. When the computer program 411 is executed by a processor, the following steps are implemented:

[0172] It should be noted that in the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0173] Those skilled in the art should understand that the embodiments of the present invention can be provided as a system, a method, or a computer program product. Therefore, the present invention can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0174] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, and the combination of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementing in the process Figure 1one or more processes and / or blocks Figure 1 a system for the functions specified in one or more blocks

[0175] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction system that implements the functions specified in the process Figure 1 one or more processes and / or blocks Figure 1 the functions specified in one or more blocks

[0176] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the process Figure 1 one or more processes and / or blocks Figure 1 the functions specified in one or more blocks

[0177] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments and all changes and modifications that fall within the scope of the present invention

[0178] Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations

Claims

1. A geological core analysis full-view digital thin-section analysis system, characterized in that: The system comprises: Matrix building module, used to obtain multi-band images of core slices and establish a multi-scale image acquisition matrix; The first function construction module is used to suppress noise and enhance edge features by using a gradient operator, and to construct a three-dimensional spatial structure function in combination with a feature point distribution function; A feature fusion module, used to perform a joint transformation on the multi-scale image acquisition matrix and the three-dimensional space structure function, calculate the fusion weight by measuring the feature similarity, and generate a feature fusion matrix; The second function building module is used to define a splicing optimization function and calculate a splicing optimization result according to the reference point coordinates and the smoothing coefficient; An integrity assessment module, used to calculate the absolute value of the gradient of the splicing optimization result, and to assess the continuity of the three-dimensional structure of the core slice in combination with a structural integrity metric function; An image update module is used to dynamically adjust the learning rate and time decay coefficient based on the gradient direction of the integrity assessment function and update the image parameters until convergence; The three-dimensional image generation module is used to select key feature points of the core slice and generate a three-dimensional visualization image of the core slice through a spatial distance attenuation coefficient.

2. The geological core analysis full-view digital thin-section analysis system according to claim 1, characterized in that: The matrix building module is also used to: Acquiring the intensity distribution of the multi-band image; Determining each image matrix element according to the intensity distribution of the multi-band image; The multi-scale image acquisition matrix is ​​constructed according to the image matrix elements.

3. The geological core analysis full-view digital thin-section analysis system according to claim 2 is characterized in that: The multi-scale image acquisition matrix is ​​expressed as: ； Where M is the multi-scale image acquisition matrix, is the image matrix element, is the weight coefficient for different image channels or scales, is the image intensity at the corresponding position and band, is the wavelength parameter, is the image matrix dimension.

4. The geological core analysis full-view digital thin-section analysis system according to claim 3 is characterized in that: The first function building module is also used for: Performing a three-dimensional integration on the multi-scale image acquisition matrix and performing noise suppression using a gradient operator to obtain a corresponding integral term; The feature point distribution function is superimposed on the integral term to obtain a space construction function.

5. The geological core analysis full-view digital thin-section analysis system according to claim 4, characterized in that: The space construction function is expressed as: ； in, is the spatial structure reconstruction function, is the voxel representation of the multi-scale image acquisition matrix, is the gradient operator, is the feature point distribution function, m is the number of feature points, They are the first adjustment parameter, the second adjustment parameter and the third adjustment parameter respectively.

6. The geological core analysis full-view digital thin-section analysis system according to claim 5, characterized in that: The feature fusion module is also used for: Obtaining a spatial structure function for describing the structural integrity and three-dimensional distribution of the core slice; defining a joint transform operator for spatially aligning the multiscale image acquisition matrix with the spatial structure function; Calculate the attenuation factor based on the feature distance to generate the feature values ​​of each fused image; The feature fusion matrix is ​​generated according to the feature values ​​of each image.

7. The geological core analysis full-view digital thin-section analysis system according to claim 6, characterized in that: The second function building block is also used for: According to the reference point coordinates, calculating the corresponding spatial attenuation factor; The splicing optimization function is determined according to the spatial attenuation factor, the smoothing coefficient and the feature fusion matrix.

8. The geological core analysis full-view digital thin-section analysis system according to claim 7, characterized in that: The integrity assessment module is also used to: Obtaining the local pore volume, connected path length and fracture density of the core slice; The structural integrity metric function is determined based on the local pore volume, the connected path length and the fracture density of the core slice.

9. The geological core analysis full-view digital thin-section analysis system according to claim 7, characterized in that: The three-dimensional image generation module is also used for: Determine a corresponding distance attenuation factor according to the key feature points of the core slice and the distance attenuation coefficient; A three-dimensional visualization image of the core slice is determined according to the distance attenuation factor and the splicing optimization function value calculated by the splicing optimization function.

10. A method for full-view digital thin-section analysis of geological core analysis, characterized in that: The method comprises: Acquire multi-band images of core slices and establish a multi-scale image acquisition matrix; The gradient operator is used to suppress noise and enhance edge features, and the three-dimensional spatial structure function is constructed by combining the feature point distribution function; Performing a joint transformation on the multi-scale image acquisition matrix and the three-dimensional space structure function, calculating fusion weights by measuring feature similarity, and generating a feature fusion matrix; According to the reference point coordinates and the smoothing coefficient, a splicing optimization function is defined and the splicing optimization result is calculated; Calculating the absolute value of the gradient of the splicing optimization result, and evaluating the continuity of the three-dimensional structure of the core slice in combination with the structural integrity metric function; Based on the gradient direction of the integrity assessment function, the learning rate and time decay coefficient are dynamically adjusted to update the image parameters until convergence; The key feature points of the core slice are selected, and a three-dimensional visualization image of the core slice is generated through a spatial distance attenuation coefficient.

Images