Shale slice lamina identification and analysis method and system

Abstract

The invention discloses a shale slice lamina identification and analysis method and system, and the method comprises the steps: converting a shale slice image into a gray matrix, optimizing a global threshold value through a quantum annealing algorithm, and dynamically adjusting a local threshold value in combination with the membrane potential of a pulse neural network; wavelet decomposition is adopted to extract multi-scale features and map the multi-scale features to spatial domain pixel blocks, and global and local thresholds are fused to realize block binaryzation; splicing the binary image to generate a white pixel trend line, identifying a bright color / dark color stripe layer through filtering fitting and dynamic parameter detection, and finally counting thickness, density and distribution characteristics. According to the invention, through the synergistic effect of the global optimization capability of quantum annealing and the local adaptive characteristic of the spiking neural network, the segmentation precision of the complex lamina is significantly improved; the calculation complexity is reduced through wavelet multi-scale analysis, and real-time processing is supported; direct mapping of geological parameters and image features is achieved through trend line parameterization detection, and high-precision data support is provided for sedimentary environment reconstruction and reservoir evaluation.

Description

Technical Field

[0001] The present invention relates to the fields of geological exploration and image processing, and particularly relates to a method and system for shale thin-section lamination identification and analysis based on quantum annealing, pulsed neural network, and pixel trend line. Background Art

[0002] The accurate identification and quantitative analysis of shale thin-section laminations are key links in sedimentology research and oil and gas reservoir evaluation. Traditional methods mainly rely on manual microscopic observation or semi-automatic image processing techniques. However, due to the complexity of lamination morphology (such as inclination, bending, and bioturbation) and noise interference in thin-section images, there are significant bottlenecks in their efficiency, accuracy, and applicability. The existing technologies mainly include the following categories:

[0003] 1. Methods based on Fourier transform and principal component analysis

[0004] Such methods extract principal component features by converting thin-section images to the frequency domain and use the axial ratio to determine the development degree of laminations. However, the global frequency-domain analysis characteristics of the Fourier transform result in its inability to capture the local time-varying features of non-stationary signals (such as curved laminations), and the principal component analysis is sensitive to data distribution. Key directional information is easily lost when frequency-domain features overlap or are interfered by noise. In addition, this method can only provide a rough assessment of the lamination development degree and lacks quantitative output of spatial-domain parameters such as thickness and density.

[0005] 2. Image processing methods based on mathematical morphology

[0006] By performing morphological operations such as erosion and dilation to segment shale and sand regions, and then statistically analyzing lamination parameters. However, this method highly depends on manually set binary thresholds, and segmentation errors are easily caused in regions with uneven illumination or bioturbation; morphological filtering may amplify local noise during the denoising process, especially generating artifacts at the edges of thin layers. Experiments show that its statistical error rate for bioturbated laminations is as high as 86.67%, severely limiting its practical application value.

[0007] 3. Intelligent identification methods based on convolutional neural network (CNN)

[0008] Although CNN shows certain potential in lamination classification tasks, it depends on a large amount of labeled data, with high labeling costs and poor generalization in small-sample scenarios. The deep network structure leads to high computational complexity and is difficult to run in real time on embedded devices; in addition, the model decision-making process lacks interpretability and cannot associate lamination physical parameters (such as composition and thickness) with classification results, restricting its credibility in geological interpretation.

[0009] None of the above prior arts effectively solve the following core problems: (1) Collaborative optimization of global and local features: It is difficult for traditional threshold segmentation or morphological operations to balance the global statistical characteristics of images and the retention of local details; (2) Insufficient dynamic adaptability: Fixed thresholds or pre-trained models cannot adapt to the spatial heterogeneity of the gray-scale distribution of thin-section images; (3) Conflict between computational efficiency and accuracy: High-precision algorithms (such as CNN) rely on expensive hardware and are difficult to meet the requirements of on-site real-time analysis; (4) Lack of geological parameter mapping: Existing methods mostly focus on the recognition of lamination morphology and lack the automated output of quantitative parameters such as thickness and density.

[0010] In this context, there is an urgent need for a shale thin-section lamination analysis technology that takes into account accuracy, efficiency, and interpretability to break through the limitations of traditional methods and meet the requirements of high-resolution geological parameters for oil and gas exploration and development. Summary of the Invention

[0011] The present invention aims at the defects of the prior art and provides a method and system for identifying and analyzing shale thin-section laminations.

[0012] In order to achieve the above invention objectives, the technical solutions adopted by the present invention are as follows:

[0013] A method for identifying and analyzing shale thin-section laminations includes the following steps:

[0014] S1. Image preprocessing: Convert the original color image of the shale thin section into a grayscale image, and the conversion formula generates a grayscale matrix based on the weighted sum of the red, green, and blue channels.

[0015] S2. Quantum annealing global threshold optimization: Calculate the histogram of the grayscale image, extract the spectral energy through Fourier transform and perform non-linear scaling;

[0016] Inject random noise into the spectral energy and iteratively optimize the global threshold through the probability acceptance mechanism of quantum annealing.

[0017] S3. Pulse neural network dynamic threshold adjustment: Dynamically adjust the local threshold based on the membrane potential update rule, lower the threshold when a pulse is triggered, slowly restore the threshold when no pulse is triggered, and limit the threshold within a preset range.

[0018] S4. Wavelet multi-scale decomposition and spatial mapping: Perform wavelet decomposition on the grayscale image to obtain the low-frequency sub-band and high-frequency sub-bands; Extract the pixel blocks corresponding to the regions in the original image as processing units through block coordinate mapping.

[0019] S5. Hybrid threshold fusion and binarization: Fuse the global threshold and the local threshold according to a preset weight to generate a hybrid threshold; Perform binarization segmentation on each pixel block to generate a binary image block.

[0020] S6, Image stitching and trend line generation: Stitch all binary blocks to generate a complete binary image; count the percentage of white pixels in each row, and generate a smooth trend line through filtering and fitting.

[0021] S7, Lamina detection and statistics: Set the detection parameters for bright and dark laminae based on the trend line, and identify the peaks and valleys; convert according to the scale and count the lamina thickness, quantity, and distribution characteristics.

[0022] Further, the conversion formula in S1 is:

[0023] Gray(x,y) = 0.299·R(x,y) + 0.587·G(x,y) + 0.114·B(x,y)

[0024] where (x,y) represents the pixel coordinates, and R, G, and B represent the intensity values of the red, green, and blue channels respectively. The converted grayscale image is stored as a two-dimensional matrix.

[0025] Further, in S2, perform a fast Fourier transform on the histogram, extract the spectral energy, and apply a non-linear scaling. The formula is:

[0026] FreqSpectrum[n] = |F(hist)[n]| 0.8

[0027] Add Gaussian random noise to the spectral energy, and select the global optimal threshold θ through the quantum annealing probability acceptance mechanism. local , with the initial temperature set to 1×10 5 , and the cooling rate is 0.98.

[0028] Further, in S3, dynamically adjust the local threshold θ based on the membrane potential update rule and the triggering mechanism. global , and the formula is:

[0029]

[0030] where I(t) represents the current pixel brightness value, and V mem (t) represents the membrane potential at time t;

[0031] When a trigger pulse occurs, the threshold is lowered, and when no trigger occurs, the threshold slowly recovers. θ local is restricted within the range of [30, 110].

[0032] Further, in S4, perform a first-level Haar wavelet decomposition on the grayscale image to obtain the low-frequency sub-band LL and the high-frequency sub-bands {LH, HL, HH}; extract the corresponding 2×2 pixel blocks in the original image through block coordinate mapping as independent processing units.

[0033] Further, in S5, set the global threshold θ globalWith the local threshold θ local Fuse by weight to generate a hybrid threshold θ mix , and the formula is:

[0034] θ mix = 0.4·θ global + 0.6·θ local

[0035] Perform binary segmentation on each 2×2 pixel block to generate a binary image block.

[0036] Furthermore, the specific process of quantum annealing global threshold optimization in S2 includes:

[0037] The spectral energy is enhanced by injecting Gaussian noise, and the noise formula is:

[0038] Energy[n] = FreqSpectrum[n]·(1 + 0.2·ξ[n])

[0039] where ξ[n]~N(0,1) is the standard normal distribution noise;

[0040] The probability acceptance formula for annealing decision is:

[0041] P accept = exp(-ΔE / T)

[0042] where ΔE represents the energy difference between the candidate threshold and the current threshold, and T represents the annealing temperature;

[0043] Gradually converge to the global optimal threshold through iteration.

[0044] Furthermore, the dynamic threshold adjustment rule of the pulsed neural network in S3 is:

[0045] When a trigger pulse occurs, the threshold reduction formula is:

[0046] θ local ←θ local - 8·(1 - e -τ / 30 )

[0047] When not triggered, the threshold recovery formula is:

[0048] θ local ←θ local + 3·e -τ / 80

[0049] where τ = 15.0, which is used to control the adjustment rate.

[0050] The present invention also discloses a shale thin-section lamination identification and analysis system, which can be used to implement the above-mentioned shale thin-section lamination identification and analysis method. Specifically, it includes:

[0051] An image preprocessing module for converting an original color image of a shale thin section into a grayscale image, where the conversion generates a grayscale matrix based on the weighted sum of the red, green, and blue channels;

[0052] A quantum annealing global threshold optimization module for calculating the histogram of the grayscale image, extracting spectral energy through Fourier transform and performing non-linear scaling. After injecting random noise into the spectral energy, the global threshold is iteratively optimized through the probability acceptance mechanism of quantum annealing;

[0053] A spiking neural network dynamic threshold adjustment module for dynamically adjusting the local threshold based on the membrane potential update rule, lowering the threshold when a spike is triggered, slowly restoring the threshold when no spike is triggered, and restricting the threshold within a preset range;

[0054] A wavelet multi-scale decomposition and spatial mapping module for performing wavelet decomposition on the grayscale image to obtain a low-frequency sub-band and a high-frequency sub-band, and extracting pixel blocks corresponding to the original image in the corresponding regions through block coordinate mapping as processing units;

[0055] A hybrid threshold fusion and binarization module for fusing the global threshold and the local threshold according to a preset weight to generate a hybrid threshold, and performing binarization segmentation on each pixel block to generate binary image blocks;

[0056] An image stitching and trend line generation module for stitching all binary blocks to generate a complete binary image, counting the percentage of white pixels in each row and generating a smooth trend line through filtering and fitting;

[0057] A lamina detection and statistics module for setting detection parameters for bright and dark laminae based on the trend line to identify peaks and valleys, and calculating the lamina thickness, quantity, and distribution characteristics according to the scale conversion.

[0058] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the above-mentioned shale thin section lamina recognition and analysis method is implemented.

[0059] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the above-mentioned shale thin section lamina recognition and analysis method is implemented.

[0060] Compared with the prior art, the advantages of the present invention are:

[0061] 1. Through the synergistic effect of global threshold optimization by quantum annealing and dynamic local threshold adjustment by pulse neural network, the problem of missegmentation caused by uneven illumination, noise interference, or complex lamination morphology in traditional methods is solved, and the robustness of lamination boundary recognition is significantly improved. At the same time, the multi-scale analysis ability of wavelet decomposition effectively preserves the spatial details of the thin-section microstructure, avoiding lamination fracture or adhesion caused by resolution loss.

[0062] 2. The block-based binaryzation processing based on wavelet block mapping significantly reduces the memory occupancy of the algorithm and supports parallel computing acceleration; the lightweight model design of the pulse neural network reduces the dependence on computing resources, enabling it to be deployed on edge devices and meet the real-time requirements.

[0063] 3. Through trend line filtering fitting and dynamic parameter detection, not only can the positions and thicknesses of bright and dark laminations be accurately identified, but also key geological parameters such as lamination density and thickness distribution heterogeneity can be quantified and statistically analyzed, providing multi-dimensional data support for sedimentary environment reconstruction and reservoir evaluation.

[0064] 4. The spectrum energy non-linear scaling and Gaussian noise injection strategy improve the adaptability of the algorithm to abnormal scenarios such as bioturbation and image blur, ensuring the stability of complex geological sample analysis.

[0065] 5. The hybrid threshold fusion mechanism and trend line parameterized detection rules have clear physical meanings, supporting the transparent verification of the algorithm logic; the modular design facilitates compatibility with different imaging devices or extension to other thin-section analysis scenarios of different lithologies. Brief Description of the Drawings

[0066] Figure 1 is the flowchart of the shale thin-section lamination recognition and analysis method according to the embodiment of the present invention.

[0067] Figure 2 is the comparison diagram between the original image of lamination development and the binary image processed by the present invention. Among them, (a) is the original image, and (b) is the binary image.

[0068] Figure 3 is the comparison diagram between the original image of underdeveloped lamination and the binary image processed by the present invention. Among them, (a) is the original image, and (b) is the binary image.

[0069] Figure 4 is the lamination statistical information diagram according to the embodiment of the present invention. Among them, (a) is the result diagram obtained from the lamination development image. According to the results in the diagram, it can be obtained that there are 8 bright laminations and 7 dark laminations, and the statistical data of the bright laminations and dark laminations are shown in the diagram. (b) is the result obtained from the underdeveloped lamination, and no image with underdeveloped lamination is misidentified as having lamination development. Detailed Embodiments

[0070] To make the objectives, technical solutions and advantages of the present invention more clear and understandable, the following further describes the present invention in detail with reference to the accompanying drawings and by way of examples.

[0071] As Figure 1 shown, the shale thin - layer lamination recognition and analysis method proposed by the present invention includes the following steps:

[0072] 1. Pre - process the original shale thin - layer image to convert it into a grayscale image. The conversion formula is as shown in Equation (1):

[0073] Gray(x,y) = 0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y) (1)

[0074] Where (x,y) represents the pixel coordinates, and R, G, and B respectively represent the intensity values (range 0 - 255) of the red, green, and blue channels. The converted grayscale image is stored as a two - dimensional matrix with the same dimension as the original image, and the value range of the elements is [0,255], where 0 represents pure black and 255 represents pure white.

[0075] 2. Quantum annealing global threshold optimization

[0076] 2.1 Calculate the 256 - level histogram of the grayscale image, count the number of pixels at each gray level, and generate a 256 - dimensional vector hist.

[0077]

[0078] In the formula, δ(·) is the indicator function, which takes 1 when the condition in the parentheses is satisfied, and 0 otherwise.

[0079] 2.2 Perform a fast Fourier transform (FFT) on the histogram, extract the spectral energy and apply a non - linear scaling:

[0080] FreqSpectrum[n] = |F(hist)[n]| 0.8 (3)

[0081] Where represents the Fourier transform operator, and the exponent 0.8 is used to enhance the low - frequency components.

[0082] 2.3 Add Gaussian random noise to the spectral energy.

[0083] Energy[n] = FreqSpectrum[n]·(1 + 0.2·ξ[n]) (4) Where ξ[n]~N(0,1) is the standard normal distribution noise.

[0084] 2.4 Annealing decision: Select a new threshold through a probability acceptance mechanism:

[0085] P accept= exp(-ΔE / T) (5)

[0086] where P accept is the acceptance probability, ΔE is the candidate threshold energy difference, T is the current temperature, the initial temperature is set to 1e5, the cooling rate is 0.98, and it gradually converges to the global optimal threshold θ global .

[0087] 3. Dynamic Threshold Adjustment of Spiking Neural Network

[0088] 3.1 Membrane Potential Update:

[0089]

[0090] In the formula, V mem (t) is the membrane potential at the previous moment, I(t) is the current pixel brightness value, and 0.85 is the membrane potential decay coefficient, simulating the ion channel leakage effect.

[0091] 3.2 Threshold Adaptive Adjustment: When the membrane potential exceeds the threshold, a spike is triggered and the threshold is lowered:

[0092] θ local ← θ local - 8·(1 - e -τ / 30 ) (7)

[0093] When not triggered, the threshold slowly recovers:

[0094] θ local ← θ local + 3·e -τ / 80 (8)

[0095] The parameter τ = 15.0 controls the adjustment rate, and the local threshold θ local is restricted within the range of [30, 110] to prevent extreme values.

[0096] 4. Wavelet Multi-scale Decomposition and Spatial Mapping

[0097] 4.1 Wavelet Decomposition: Perform one-level Haar wavelet decomposition on the grayscale image to obtain the low-frequency sub-band LL and the high-frequency sub-bands {LH, HL, HH}. After decomposition, the size of the low-frequency sub-band LL is 1 / 2 of the original image, that is, each low-frequency coefficient corresponds to a 2×2 pixel block in the original image.

[0098] 4.2 Block Coordinate Mapping: For each position (i, j) in the low-frequency sub-band, calculate the starting coordinates of the corresponding 2×2 pixel block in the original image:

[0099] x start , y start . This mapping relationship stems from the downsampling characteristic of wavelet decomposition, and each low-frequency coefficient corresponds to the original Figure 2 ×2 area.

[0100] 4.3 Local block extraction: Extract the corresponding block data from the original grayscale matrix:

[0101] Block = gray_array[x start :x start +2, y start :y start +2] (9)

[0102] where Block is a 2×2 pixel block and gray_array is the grayscale matrix. This operation treats each 2×2 pixel block as an independent processing unit.

[0103] 5. Hybrid threshold fusion and binarization

[0104] 5.1 Fuse the global threshold θ obtained by quantum annealing global with the local threshold θ generated by SNN local by weight:

[0105] θ mix = 0.4·θ global + 0.6·θ local (10)

[0106] The weight coefficient is determined by cross-validation, giving priority to retaining local detail features.

[0107] 5.2 Binarization decision: Perform threshold segmentation on each 2×2 pixel block:

[0108]

[0109] where BinaryBlock[m, n] represents the pixel value at the m-th row and n-th column in the 2×2 pixel block after binarization, and Block[m, n] represents the pixel grayscale value at the m-th row and n-th column within the block (m, n ∈ {0, 1})

[0110] 6. Stitch all the processed 2×2 binary blocks according to the coordinate mapping relationship to generate a complete binary image.

[0111] 7. Count the number of white pixels in the binarized image row by row to generate a white pixel percentage curve graph.

[0112] 8. Apply Savitzky-Golay filtering to the white pixel percentage curve graph, and then perform a linear fit on the white pixel percentage curve using a first-order polynomial to generate a trend line. The parameter settings are as follows:

[0113] window_size = min(51, len(profile) / / 3 * 2 - 1)

[0114] trend_curve = savgol_filter(profile, window_size, 1)

[0115] Among them, window_size is the window size of the Savitzky-Golay filter, profile is a 1D array representing the brightness percentage of each row, trend_curve represents the generated trend line, and savgol_filter is the Savitzky-Golay filter.

[0116] 9. Apply Gaussian filtering to the trend line to smooth the original curve and retain the macroscopic trend. The parameter settings are as follows:

[0117] trend_curve = gaussian_filter1d(trend_curve, sigma = 5)

[0118] Among them, gaussian_filter1d is the Gaussian filter, and sigma = 5 is the smoothing intensity.

[0119] 10. Combine the original shale thin-section image, set the feature detection parameters for light-colored laminae and dark-colored laminae, and then identify the light-colored laminae and dark-colored laminae.

[0120] 10.1 The feature detection parameter settings for light-colored laminae are as follows:

[0121] 'bright': {'height': 10, 'prominence': 3, 'width': 5, 'distance': 20}

[0122] Among them, bright represents the light-colored laminae, height represents the lowest peak height threshold, prominence represents the minimum peak prominence, width represents the minimum peak width, and distance represents the minimum adjacent peak interval.

[0123] 10.2 The feature detection parameter settings for dark-colored laminae are as follows:

[0124] 'dark': {'height': np.percentile(-trend_curve, 50), 'prominence': 3, 'width': 5,

[0125] 'distance': 20}

[0126] Among them, dark represents the dark stripe layer, height represents the minimum height threshold of the wave valley, np.percentile(-trend_curve, 50) represents the dynamic height threshold (taking the median), prominence represents the minimum prominence of the wave valley, width represents the minimum width of the wave valley, and distance represents the minimum interval between adjacent wave valleys.

[0127] 10.3 Directly find the peaks corresponding to the bright stripe layers on the trend line, and find the valleys corresponding to the dark stripe layers on the trend line.

[0128] 11 According to the scale conversion, count the thickness of each stripe layer, the maximum thickness and the minimum thickness of the stripe layer. Count the number of bright stripe layers and dark stripe layers, and the total number of stripe layers.

[0129] The present invention designs a binarization method based on quantum annealing and pulsed neural network for shale thin-section images, which is characterized in that: a globally optimized threshold is generated by a quantum annealing algorithm with Fourier spectrum enhancement, the local threshold is dynamically adjusted by combining a pulsed neural network in the wavelet domain, and a hybrid weighting strategy is used to fuse global and local features; among them, the quantum annealing process realizes the energy barrier crossing through a Hamiltonian model injecting golden ratio noise, the pulsed neural network adopts a time-varying membrane potential update rule to adaptively adjust the threshold range, and finally a binary image is obtained by covering the whole image through a block hybrid threshold mapping of Haar wavelet decomposition. Based on this binary image, a curve of the proportion of white pixels is generated. First, Savitzky-Golay filtering is used, and the key parameter is set as window_size = min(51, len(profile) / / 3 * 2 - 1); then a first-order polynomial is used to fit the curve of the proportion of white pixels to generate a trend line of the proportion of white pixels; then Gaussian filtering is used in combination with the actual shale thin-section image, and the key parameter is set as trend_curve = gaussian_filter1d(trend_curve, sigma = 5); finally, the detection parameters of bright stripe layers and dark stripe layers are set based on the trend line of the proportion of white pixels to find bright stripe layers, dark stripe layers and perform statistical analysis, and the key parameters are set as 'bright': {'height': 10, 'prominence': 3, 'width': 5, 'distance': 20}, 'dark': {'height': np.percentile(-trend_curve, 50), 'prominence': 3, 'width': 5, 'distance': 20}.

[0130] Such as Figures 2 to 4As shown in the figure, the result of shale lamina identification and analysis using the method of the present invention shows that: a binary image that conforms to the original shale thin-section image can be obtained, the lamina structure is more obvious, and the connectivity of the laminae is better ( Figure 2 ), the bright pixels are more prominent ( Figure 3 ), the positions and widths of the bright and dark laminae can be directly seen according to the trend line, and the lamina statistical information of each thin section can be obtained extremely conveniently ( Figure 4 ) In thin sections where the laminae are not well developed, no laminae are misidentified.

[0131] In another embodiment of the present invention, a shale thin-section lamina identification and analysis system is provided. This system can be used to implement the above-mentioned shale thin-section lamina identification and analysis method. Specifically, it includes:

[0132] An image preprocessing module for converting the original color image of the shale thin section into a grayscale image, and the conversion generates a grayscale matrix based on the weighted sum of the red, green, and blue channels;

[0133] A quantum annealing global threshold optimization module for calculating the histogram of the grayscale image, extracting the spectral energy through Fourier transform and performing non-linear scaling. After injecting random noise into the spectral energy, the global threshold is iteratively optimized through the probability acceptance mechanism of quantum annealing;

[0134] A pulsed neural network dynamic threshold adjustment module for dynamically adjusting the local threshold based on the membrane potential update rule, lowering the threshold when a pulse is triggered, slowly restoring the threshold when no pulse is triggered, and restricting the threshold within a preset range;

[0135] A wavelet multi-scale decomposition and spatial mapping module for performing wavelet decomposition on the grayscale image to obtain a low-frequency sub-band and a high-frequency sub-band, and extracting pixel blocks corresponding to the original image in the corresponding area through block coordinate mapping as processing units;

[0136] A hybrid threshold fusion and binarization module for fusing the global threshold and the local threshold according to a preset weight to generate a hybrid threshold, and performing binarization segmentation on each pixel block to generate a binary image block;

[0137] An image stitching and trend line generation module for stitching all binary blocks to generate a complete binary image, counting the percentage of white pixels in each row and generating a smooth trend line through filtering and fitting;

[0138] A lamina detection and statistics module for setting detection parameters for bright and dark laminae based on the trend line to identify peaks and valleys, and calculating the lamina thickness, quantity, and distribution characteristics according to the scale conversion.

[0139] In another embodiment of the present invention, a terminal device is provided. The terminal device includes a processor and a memory. The memory is used to store a computer program, and the computer program includes program instructions. The processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding function. The processor described in the embodiments of the present invention can be used for the operation of the shale thin-section lamination identification and analysis method.

[0140] In another embodiment of the present invention, a storage medium is also provided, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is the memory device in the terminal device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, and the operating system of the terminal is stored in this storage space. And, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory.

[0141] One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the shale thin-section lamination identification and analysis method in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor.

[0142] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0143] 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 flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, 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 processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0144] 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 generate a manufactured article including instruction means that implement the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0145] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0146] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the implementation methods of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. A method for identifying and analyzing shale thin - layer laminations, characterized in that, It includes the following steps: S1. Image preprocessing: Convert the original color image of the shale thin section into a grayscale image, and the conversion formula generates a grayscale matrix based on the weighted sum of the red, green, and blue channels; S2. Quantum annealing global threshold optimization: Calculate the histogram of the grayscale image, extract the spectral energy through Fourier transform and perform non-linear scaling; Inject random noise into the spectral energy, and iteratively optimize the global threshold through the probability acceptance mechanism of quantum annealing; S3. Pulse neural network dynamic threshold adjustment: Dynamically adjust the local threshold based on the membrane potential update rule, lower the threshold when a pulse is triggered, slowly recover the threshold when not triggered, and limit the threshold within a preset range; S4. Wavelet multi-scale decomposition and spatial mapping: Perform wavelet decomposition on the grayscale image to obtain the low-frequency sub-band and high-frequency sub-bands; Extract the pixel blocks corresponding to the regions in the original image through block coordinate mapping as processing units; S5. Hybrid threshold fusion and binarization: Fuse the global threshold and the local threshold according to a preset weight to generate a hybrid threshold; Perform binarization segmentation on each pixel block to generate binary image blocks; S6. Image stitching and trend line generation: Stitch all binary blocks to generate a complete binary image; Statistically calculate the percentage of white pixels in each row, and generate a smooth trend line through filtering and fitting; S7. Lamina detection and statistics: Set the detection parameters for bright and dark laminae based on the trend line, and identify the peaks and valleys; Calculate and statistically analyze the lamina thickness, quantity, and distribution characteristics according to the scale conversion.

2. The method for identifying and analyzing the laminae of a shale thin section according to claim 1, wherein: The conversion formula in S1 is: Gray(x,y)=0.299·R(x,y)+0.587·G(x,y)+0.114·B(x,y) where (x,y) represents the pixel coordinates, and R, G, and B respectively represent the intensity values of the red, green, and blue channels; The converted grayscale image is stored as a two-dimensional matrix.

3. The method for identifying and analyzing the laminae of a shale thin section according to claim 1, wherein: In S2, perform a fast Fourier transform on the histogram, extract the spectral energy and apply non-linear scaling, and the formula is: FreqSpectrum[n] = |F(hist)[n]| 0.8 Add Gaussian random noise to the spectral energy and select the global optimal threshold θ through the quantum annealing probability acceptance mechanism local , set the initial temperature to 1×10 5 , and the cooling rate is 0.

98.

4. The method for identifying and analyzing the laminae of a shale thin section according to claim 1, wherein: Dynamically adjust the local threshold θ based on the membrane potential update rule and triggering mechanism in S3 global , and the formula is: Among them, I(t) represents the current pixel brightness value, and V mem (t) represents the membrane potential at time t; The threshold is lowered when the trigger pulse is present and slowly recovers when not triggered, θ local is restricted within the range of [30, 110].

5. The method for identifying and analyzing the laminae of a shale thin section according to claim 1, wherein: In S4, perform a first-level Haar wavelet decomposition on the grayscale image to obtain the low-frequency sub-band LL and the high-frequency sub-bands {LH, HL, HH}; Extract the corresponding 2×2 pixel blocks in the original image through block coordinate mapping as independent processing units.

6. The shale thin - layer lamination recognition and analysis method according to claim 1, characterized in that: In S5, the global threshold θ global is fused with the local threshold θ local by weights to generate a mixed threshold θ mix , and the formula is: θ mix = 0.4·θ global + 0.6·θ global Perform binarization segmentation on each 2×2 pixel block to generate binary image blocks.

7. The method for identifying and analyzing the laminae of a shale thin section according to claim 1, wherein: The specific process of quantum annealing global threshold optimization in S2 includes: The spectral energy is enhanced in robustness by injecting Gaussian noise, and the noise formula is: Energy[n]=FreqSpectrum[n]·(1+0.2·ξ[n]) where ξ[n]~N(0,1) is the standard normal distribution noise; The probability acceptance formula for the annealing decision is: P accept = exp(-ΔE / T) where ΔE represents the energy difference between the candidate threshold and the current threshold, and T represents the annealing temperature; It gradually converges to the global optimal threshold through iteration.

8. The shale thin-layer lamination identification and analysis method according to claim 1, characterized in that: The dynamic threshold adjustment rule of the spiking neural network in S3 is as follows: When a pulse is triggered, the threshold reduction formula is: θ local ←θ local -8·(1 - e -τ / 30 ) When no pulse is triggered, the threshold recovery formula is: θ local ←θ local +3·e -τ / 80 where τ = 15.0, which is used to control the adjustment rate.

9. A shale thin-section lamination identification and analysis system, characterized in that: This system can be used to implement the shale thin - layer lamination identification and analysis method described in any one of claims 1 to 8. Specifically, it includes: An image pre - processing module for converting the original color image of the shale thin slice into a grayscale image. The conversion is based on the weighted sum of the red, green, and blue channels to generate a grayscale matrix; A quantum annealing global threshold optimization module for calculating the histogram of the grayscale image, extracting the spectral energy through Fourier transform and performing non - linear scaling. After injecting random noise into the spectral energy, it iteratively optimizes the global threshold through the probability acceptance mechanism of quantum annealing; A spiking neural network dynamic threshold adjustment module for dynamically adjusting the local threshold based on the membrane potential update rule, reducing the threshold when a pulse is triggered, slowly recovering the threshold when no pulse is triggered, and restricting the threshold within a preset range; A wavelet multi - scale decomposition and spatial mapping module for performing wavelet decomposition on the grayscale image to obtain low - frequency and high - frequency sub - bands, and extracting pixel blocks corresponding to the original image in the corresponding regions through block - coordinate mapping as processing units; A hybrid threshold fusion and binarization module for fusing the global threshold and the local threshold according to a preset weight to generate a hybrid threshold, and performing binarization segmentation on each pixel block to generate binarized image blocks; An image stitching and trend - line generation module for stitching all binarized blocks to generate a complete binarized image, counting the percentage of white pixels in each row and generating a smooth trend - line through filtering and fitting; A lamination detection and statistics module for setting detection parameters for bright and dark laminations based on the trend - line to identify peaks and valleys, and calculating the lamination thickness, quantity, and distribution characteristics according to the scale conversion; 10. A computer-readable storage medium, characterized in that: It stores a computer program, and when the program is executed by a processor, it implements the shale thin - layer lamination identification and analysis method described in any one of claims 1 to 8.

Images