Rock damage quantification method based on SEM image recognition and application

Abstract

The application discloses a rock damage quantification method based on SEM electron microscope image recognition and application, and belongs to the technical field of rock material micro-characterization and intelligent recognition. In view of the problems that the existing rock damage quantification methods are generally out of touch with micro-evidence, are difficult to quickly review, and are insufficient in sample comparison ability, etc., the application extracts micro-damage characteristic parameters of a rock sample to be measured through an SEM electron microscope image, and then constructs a mapping relationship between the micro-damage characteristic parameters and an equivalent damage variable, so that unified dimension characterization and quantitative evaluation of the rock micro-damage state are realized. In addition, the application also determines the weight coefficients of different micro-damage characteristic parameters in the equivalent damage variable by minimizing the difference between the damage variable calculated by real measurement and the equivalent damage variable, so that the calculation accuracy of the equivalent damage variable is significantly improved, and a reliable intelligent technical means is provided for rock mass engineering and material damage analysis.

Description

Technical Field

[0001] This invention relates to the field of microscopic characterization and intelligent identification technology of rock materials, and in particular to a method and application for quantifying rock damage based on SEM electron microscopy image recognition. Background Technology

[0002] Rock materials undergo irreversible microstructural degradation under loading, unloading, cyclic fatigue, freeze-thaw cycles, alternating wet and dry conditions, and chemical erosion. This degradation manifests as microcrack initiation and propagation, evolution of pore morphology and connectivity, debonding of grain boundaries, and weakening of cementation. The accumulation of this micro-damage leads to a continuous decline in macroscopic parameters such as elastic modulus, strength, wave velocity, and permeability, ultimately affecting the stability and service safety of the rock mass. Therefore, establishing an efficient, objective, quantifiable, comparable, and accurate method for assessing the degree of rock damage is a key technical requirement in rock mechanics testing and engineering safety assessment.

[0003] Current rock damage assessment methods largely rely on macroscopic mechanical indices, nondestructive testing signals, or damage variables constructed from rock constitutive inversion. However, macroscopic indices are insensitive to early micro-damage and struggle to reveal mechanisms; signal-based methods are susceptible to coupling conditions and noise interference, resulting in insufficient comparability across experiments; and inversion-based methods are highly dependent on model assumptions and parameter identification, making it difficult to maintain stable and consistent results under complex microstructural evolution conditions. These methods generally suffer from problems such as a disconnect between damage quantification and microscopic evidence, difficulty in rapid verification, and insufficient sample comparison capabilities.

[0004] Scanning electron microscopy (SEM) can provide high-resolution information on the microscopic morphology of rock surfaces, theoretically directly characterizing damage features such as microcracks and pores, providing intuitive evidence for damage quantification. However, in practical applications, SEM images often exhibit characteristics such as uneven grayscale, noise interference, complex textures, large variations in mineral phase contrast, and significant differences at different magnification scales. This leads to insufficient stability in traditional methods such as threshold segmentation, edge detection, and morphological processing, sensitivity to parameters and imaging conditions, and difficulty in ensuring the continuity of the slender topology of cracks and the accuracy of pore boundaries. Furthermore, existing SEM applications mostly remain at the level of qualitative interpretation or limited manual statistics, lacking a general recognition framework for SEM images and a quantitative link to further efficiently and accurately invert observable microstructural features into damage variables / damage levels. Therefore, it is difficult to achieve the engineering goal of quickly and reliably outputting the degree of rock damage solely through SEM scanning. Summary of the Invention

[0005] To address the aforementioned problems, this invention aims to provide a method and application for quantifying rock damage based on SEM (Semiconductor Electron Microscopy) image recognition. It extracts microscopic damage characteristic parameters from the rock sample under test using SEM images, and then establishes a mapping relationship between these parameters and equivalent damage variables, achieving a unified dimensional characterization and quantitative evaluation of the rock's microscopic damage state. When determining the weighting coefficients for different microscopic damage characteristic parameters, the damage variable of the corresponding lithological sample is measured and calculated, and compared with the equivalent damage variable calculated based on the SEM image. Minimizing the difference between the damage variable and the equivalent damage variable, and optimizing the weighting coefficients of different microscopic damage characteristic parameters in the equivalent damage variable, significantly improves the accuracy of the equivalent damage variable calculation.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: On one hand, the present invention provides a method for quantifying rock damage based on SEM electron microscopy image recognition, the method comprising: SEM images of the rock sample under different magnification conditions and corresponding imaging metadata were acquired, and the SEM images were preprocessed based on the imaging metadata. Damage features of the rock sample were identified based on the preprocessed SEM images of the rock sample. Based on the damage feature identification results, the microscopic damage feature parameters of the rock sample to be tested are extracted. Calculate the membership function value for each micro-damage characteristic parameter. ; Based on the membership function values ​​of different micro-damage characteristic parameters The equivalent damage variables of the rock sample under test are calculated by taking the weighting coefficients corresponding to different micro-damage characteristic parameters and classifying the damage level of the rock sample under test.

[0007] Furthermore, the microscopic damage characteristic parameters of the rock sample to be tested include: fracture length density L d , percentage of fracture area A c C, the parameter of fracture connectivity n Porosity and texture entropy T e .

[0008] Furthermore, the membership function values ​​of the micro-damage characteristic parameters The calculation method is as follows: ; In the formula, x represents the calculated value of the microscopic damage characteristic parameter, a S and b S Let S be the two segmented thresholds for the microscopic damage characteristic parameter S, and , This represents the membership function value corresponding to the calculated value x of the micro-damage characteristic parameter S; the micro-damage characteristic parameter S includes L d A c C n , and T e .

[0009] Furthermore, the equivalent damage variable of the rock sample to be tested The calculation method is as follows: ; In the formula, Let be the weighting coefficients corresponding to the microscopic damage feature parameter S, and n represent the number of microscopic damage feature parameters S. When S = 1, 2, 3, 4, 5, they correspond to L respectively. d A c C n , and T e .

[0010] Furthermore, the method for determining the weighting coefficients corresponding to the microscopic damage characteristic parameter S is as follows: SEM images of lithological samples corresponding to the rock sample under test at different damage evolution stages were obtained and preprocessed. At the same time, the damage variable D corresponding to the lithological samples at different damage evolution stages was obtained. Based on the preprocessed SEM electron microscope images of the lithological samples, damage features of the lithological samples are identified, and microscopic damage feature parameters of the lithological samples are extracted. Based on the membership functions of different microscopic damage characteristic parameters, the equivalent damage variables of lithological samples are calculated. ; Equivalent damage variables based on lithological samples The weighting coefficients corresponding to the micro-damage characteristic parameter S are optimized based on the damage variable D corresponding to the lithological sample.

[0011] Furthermore, the method for calculating the damage variable D corresponding to the lithological sample is as follows: ; In the formula, E0 is the elastic modulus in the undamaged state, and E is the elastic modulus of the lithological sample at the damage evolution stage.

[0012] Furthermore, the objective function for optimizing the weighting coefficients corresponding to the micro-damage characteristic parameter S is: ; In the formula, min represents the minimum value. It is an absolute value.

[0013] Furthermore, preprocessing of SEM electron microscope images based on imaging metadata includes: Scale calibration and unification of SEM electron microscope images based on imaging metadata; Gray-level normalization and noise suppression were performed on the scale-calibrated and unified SEM electron microscope images; Local contrast enhancement and feature saliency enhancement were performed on the noise-suppressed SEM images; The SEM electron microscope images after local contrast enhancement and feature saliency processing are cropped and scaled for output.

[0014] On the other hand, the present invention also provides a rock damage quantification system based on SEM electron microscopy image recognition, for implementing the method described above, the system comprising: The SEM image acquisition and preprocessing module is used to acquire SEM images of the rock sample under different magnification conditions and the corresponding imaging metadata, and to preprocess the SEM images based on the imaging metadata. The damage feature recognition module identifies damage features of the rock sample based on the preprocessed SEM electron microscope image of the rock sample to be tested. The micro-damage feature parameter extraction module extracts the micro-damage feature parameters of the rock sample to be tested based on the damage feature identification results. The rock damage quantification module calculates the membership function value of each micro-damage characteristic parameter based on the micro-damage characteristic parameters, and then calculates the equivalent damage variable of the rock sample to be tested, and classifies the damage level of the rock sample to be tested.

[0015] In another aspect, the present invention also provides an electronic device, the device including at least one processor; and a memory communicatively connected to the processor; wherein the memory stores instructions executed by the processor to enable the processor to perform the method described above.

[0016] The beneficial effects of this invention are: 1. The rock damage quantification method based on SEM image recognition in this invention extracts microscopic damage feature parameters of the rock sample under test from SEM images, and then constructs a mapping relationship between the microscopic damage feature parameters and equivalent damage variables. Different microscopic damage feature parameters are converted into membership function values ​​with unified dimensions. The equivalent damage variable is then calculated by weighted summation of the membership function values, realizing a unified dimensional characterization and quantitative evaluation of the rock microscopic damage state. When extracting the microscopic damage feature parameters of the rock sample under test, the scale is calibrated by using SEM images with different magnifications before extracting the microscopic damage feature parameters. This can fully explore the multi-scale information of cracks, pores and rock matrix structure in the SEM images, significantly improving the objectivity and consistency of the quantitative identification of damage in the rock sample under test.

[0017] 2. The rock damage quantification method based on SEM electron microscopy image recognition in this invention first calculates the damage variables of the lithological sample corresponding to the rock sample under test at different damage evolution stages based on the elastic modulus test, minimizes the difference between the damage variable and the equivalent damage variable, and optimizes the weight coefficients of different micro-damage characteristic parameters in the calculation process of the equivalent damage variable. This can significantly improve the calculation accuracy of the equivalent damage variable and provide a reliable intelligent technical means for rock engineering and material damage analysis. Attached Figure Description

[0018] Figure 1 This is a flowchart of the rock damage quantification method based on SEM electron microscopy image recognition in this invention; Figure 2 This is a flowchart of the preprocessing operation of SEM electron microscope images based on imaging metadata in this invention; Figure 3 This is a flowchart illustrating the operation of damage feature identification and microscopic damage feature parameter extraction in this invention; Figure 4 This is a flowchart illustrating the operation of calculating the equivalent damage variable of the rock sample under test based on microscopic damage characteristic parameters and classifying the damage level of the rock sample under test in this invention. Figure 5 These are SEM images of four typical engineering lithological samples—granite, sandstone, limestone, and basalt—from Example 2 of this invention. Figure 6 This is the identification result of cracks and pores in granite at different damage evolution stages in Embodiment 2 of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0020] Example 1: Example 1 provides a method for quantifying rock damage based on SEM electron microscopy image recognition. This method extracts microscopic damage characteristic parameters of the rock sample under test from SEM electron microscopy images, and then constructs a mapping relationship between microscopic damage characteristic parameters and equivalent damage variables. Different microscopic damage characteristic parameters are converted into membership function values ​​with unified dimensions. Then, the equivalent damage variable is obtained by weighted summation of the membership function values, realizing a unified dimensional characterization and quantitative evaluation of the microscopic damage state of rocks.

[0021] Specifically, refer to the appendix Figure 1 As shown, the method includes the following steps: Step 1: Obtain SEM images of the rock sample under different magnification conditions and the corresponding imaging metadata, and preprocess the SEM images based on the imaging metadata. More specifically, by obtaining SEM images of the rock sample under different magnifications, we can obtain microstructure images of the rock sample under different magnifications.

[0022] Imaging metadata includes magnification and pixel resolution. The imaging metadata corresponding to SEM images under different magnification conditions is different. Based on the imaging metadata, the scale of SEM images under different magnification conditions can be unified to ensure that the micro-damage feature parameters extracted by the micro-damage feature parameter extraction module are comparable.

[0023] See attached document Figure 2 As shown, the specific process of preprocessing SEM electron microscope images based on imaging metadata includes the following sub-steps: Sub-step 101: Scale calibration and unification of SEM electron microscope images based on imaging metadata; Establish the correspondence between pixel length and actual length of SEM images under different magnification conditions to form pixel-length calibration coefficients; and unify SEM images under different magnification conditions to a preset target scale based on the pixel-length calibration coefficients.

[0024] Sub-step 102: Perform grayscale normalization and noise suppression processing on the scale-calibrated and unified SEM electron microscope images; Gray-level normalization was performed on the SEM images to map the image gray-level range to a uniform gray-level interval. Noise suppression was then performed on the normalized SEM images, using edge-preserving filtering to remove speckle noise and high-frequency noise, thus maintaining the continuity of microcrack boundaries and pore boundaries.

[0025] Sub-step 103: Perform local contrast enhancement and feature saliency processing on the noise-suppressed SEM electron microscope image; Local contrast enhancement processing was performed on the noise-suppressed SEM images to make the gray-level differences between microcracks, pores and matrix regions more significant; morphological enhancement processing was performed on the SEM images after local contrast enhancement to improve the segmentability and connectivity preservation of microcracks.

[0026] Sub-step 104: Perform field-of-view cropping and scale-unified output on the SEM electron microscope image after local contrast enhancement and feature saliency processing; The field of view of the SEM electron microscope images after local contrast enhancement and feature saliency processing is cropped and aligned to ensure that the SEM electron microscope images under different magnification conditions have a consistent field of view scale and statistical aperture, providing a consistent basis for subsequent extraction of microscopic damage feature parameters.

[0027] Step 2: Based on the preprocessed SEM images of the rock sample to be tested, damage features of the rock sample to be tested are identified. This step utilizes image recognition technology to identify damage characteristics (including porosity and fractures) of the pre-processed SEM images of the rock sample. Damage regions are segmented from the pre-processed SEM images to identify fractures and pores, and the fracture centerline framework and pore objects are extracted. This allows the microscopic damage feature parameter extraction module to extract the microscopic damage feature parameters. The overall flowchart is attached. Figure 3 As shown.

[0028] In some embodiments, step 2 specifically includes the following sub-steps: Sub-step 201: Image segmentation; Damage regions were segmented from the pre-processed SEM images of the rock samples to be tested, distinguishing between the rock matrix region, pore region and microcrack region, and outputting pore binary mask images and crack binary mask images.

[0029] Sub-step 202: Crack feature identification; Skeletonization is performed on the binary mask image of the fracture to obtain the fracture centerline skeleton; connected component analysis is performed on the fracture centerline skeleton to extract fracture segments, fracture nodes and fracture branches, forming a set of fracture topological structures to characterize the geometric morphology and connectivity features of the fracture. Sub-step 203: Pore feature identification; Connected component segmentation is performed on the pore binary mask image to extract pore objects, forming a pore object set. Statistical analysis is then performed on the number, area, and spatial distribution characteristics of the pore objects to characterize the pore damage features of the rock.

[0030] Step 3: Based on the damage feature identification results, extract the microscopic damage feature parameters of the rock sample to be tested; Specifically, the micro-damage characteristic parameter extraction module extracts the following micro-damage characteristic parameters from the rock sample under test: fracture length density L d , percentage of fracture area A c C, the parameter of fracture connectivity n Porosity and texture entropy T e .

[0031] Among them, the crack length density L d The calculation formula is based on the fracture centerline skeleton and the field of view area. ; In the formula, l i Let N be the length of the skeleton of the centerline of the i-th fracture, A be the field of view corresponding to the SEM image of the rock sample to be tested, and N be the length of the skeleton of the centerline of the i-th fracture. l denoted as the number of fracture centerline skeletons, and i is the index for counting fracture centerline skeletons.

[0032] Crack area percentage A c The calculation formula is derived from the area of ​​the slit mask and the field of view: ; In the formula, The area is the crack area.

[0033] C-type fissure connectivity parameter n It is calculated based on the ratio of the pixel area of ​​the largest connected region in the crack topology to the total pixel area of ​​the crack, and its expression is: ; In the formula, This represents the total number of pixels contained in the largest gap connected region. This represents the total number of pixels in all fractured regions of the image. Connected component identification employs a connected component labeling algorithm based on binary images, with 8-neighborhood determination being the preferred method for topological clustering.

[0034] Porosity The calculation formula is derived from the area of ​​the aperture mask and the field of view: ; In the formula, This indicates the area of ​​the pore mask.

[0035] Extract gray-level co-occurrence matrix texture features within the masked area of ​​the rock matrix region and calculate texture entropy T. e The calculation formula is: ; In the formula, This represents the joint probability of gray level m and gray level n in the gray-level co-occurrence matrix under given spatial orientation and pixel spacing; m represents the gray level index of the first pixel; n represents the gray level index of its neighboring pixels; Q represents the total number of quantized gray levels.

[0036] Step 4: Calculate the membership function value for each microscopic damage characteristic parameter. ; The membership function value for each microscopic damage characteristic parameter is calculated as follows: ; In the formula, x represents the calculated value of the microscopic damage characteristic parameter. This represents the membership function value corresponding to the calculated value x of the micro-damage characteristic parameter S, a S and b S Let S be the two segmented thresholds for the microscopic damage characteristic parameter S, and The microscopic damage characteristic parameter S includes the crack length density L. d , percentage of fracture area A c C, the parameter of fracture connectivity n Porosity and texture entropy T e .

[0037] For different micro-damage feature parameters, the membership function value can map the calculated value of the micro-damage feature parameter to the range of [0,1], thereby forming a unified mapping channel of multi-scale micro-damage feature parameter → membership vector.

[0038] Step 5: Based on the membership function values ​​of different microscopic damage characteristic parameters The equivalent damage variables of the rock sample under test are calculated by taking the weighting coefficients corresponding to different micro-damage characteristic parameters and classifying the damage level of the rock sample under test.

[0039] In this invention, the equivalent damage variable is calculated as follows: ; In the formula, The equivalent damage variable for the rock sample to be tested is... These are the weighting coefficients corresponding to the microscopic damage characteristic parameter S. Let be the membership function value of the micro-damage characteristic parameter S, and n represent the number of micro-damage characteristic parameters S. In this embodiment, when n=5 and S=1,2,3,4,5, they correspond to the crack length density L, respectively. d , percentage of fracture area A c C, the parameter of fracture connectivity n Porosity and texture entropy T e .

[0040] In this invention, the equivalent damage variable of the rock sample to be tested is calculated. Classify rock damage levels. Equivalent damage variables. The value range is [0,1], with a larger value indicating a higher level of rock damage. Refer to the appendix for the overall process of steps 4 and 5. Figure 4 As shown.

[0041] In some embodiments, rock damage levels are classified into five categories: no damage, minor damage, moderate damage, moderate damage, and severe damage, with corresponding equivalent damage variable ranges as follows: No damage: <D1; Minor injury: D1≤ <D2; Moderate damage: D2≤ <D3; More severe injury: D3≤ <D4; Severe injury: ≥D4; Wherein, D1~D4 are the equivalent damage variable thresholds corresponding to different rock damage grades. When classifying rock damage grades for different lithologies, the equivalent damage variable thresholds can be determined based on calibration samples. For example, in this embodiment, taking sandstone as an example, the equivalent damage variable thresholds corresponding to different rock damage grades are shown below: No damage: < 0.10; Minor damage: 0.10 ≤ < 0.30; Moderate damage: 0.30 ≤ < 0.50; More severe injury: 0.50 ≤ < 0.70; Severe injury: ≥ 0.70.

[0042] Example 2: Example 2 provides a method for determining the weighting coefficients corresponding to the microscopic damage characteristic parameters described in Example 1, specifically including the following steps: Step 1: Obtain SEM images of lithological samples corresponding to the rock sample to be tested at different damage evolution stages and perform preprocessing. At the same time, conduct elastic modulus tests on lithological samples at different damage evolution stages to obtain the damage variable D corresponding to the lithological samples at different damage evolution stages.

[0043] Among them, the damage variable D is determined using the continuous damage mechanics definition based on elastic modulus degradation, and the calculation method is as follows: ; In the formula, E0 is the elastic modulus in the undamaged state, and E is the elastic modulus of the lithological sample at the corresponding damage evolution stage.

[0044] In the embodiments of this invention, four typical engineering lithological samples—granite, sandstone, limestone, and basalt—were selected as the research objects. For each lithological sample, five damage evolution stages (no damage, slight damage, moderate damage, relatively heavy damage, and severe damage) were set. At each damage evolution stage, ten SEM images of the lithological samples under different magnification conditions were acquired. Some of the SEM images are attached. Figure 5 As shown in Table 1 below, multiple sets of elastic modulus tests were conducted to form a damage calibration dataset containing four types of lithology, twenty damage evolution stages, a total of two hundred SEM images, and seventeen sets of elastic modulus tests. This dataset is used to support the training of the rock damage prediction model that integrates multi-scale damage feature parameters in this invention.

[0045] Table 1 Damage calibration dataset:

[0046] Obtain imaging metadata corresponding to the SEM electron microscope image, and preprocess the SEM electron microscope image based on the imaging metadata. The preprocessing method is the same as the preprocessing method of the SEM electron microscope image acquisition and preprocessing module in Example 1.

[0047] The average value of the elastic modulus measured at different damage evolution stages of the same lithological sample was taken to obtain the initial elastic modulus and the current elastic modulus corresponding to different damage evolution stages, and a damage calibration sample set for different lithological samples was constructed, as shown in Table 2 below.

[0048] Table 2 Damage calibration sample sets for samples of different lithologies ; Step 2: Based on the preprocessed SEM images of the lithological samples, damage features are identified, and microscopic damage characteristic parameters are extracted. For example, the identification results of fractures and pores in granite are shown in the attached figure. Figure 6 As shown, (a) is an SEM image of granite at the undamaged or slightly damaged stage and the results of crack and pore identification; (b) is an SEM image of granite at the moderately damaged stage and the results of crack and pore identification; and (c) is an SEM image of granite at the heavily damaged or severely damaged stage and the results of crack and pore identification.

[0049] All three sets of SEM images were acquired under the same imaging conditions: magnification of 2000 x, accelerating voltage of 5 kV, working distance of 6.8 mm, and pixel resolution of 2048×2048. They were used to compare the evolutionary differences in crack connectivity, pore morphology, and spatial correlation characteristics at different damage stages.

[0050] The SEM images shown in each figure were acquired under the same magnification conditions to compare the morphology, distribution, and connectivity changes of microcracks and pores at different damage evolution stages.

[0051] The specific operation process of step two is the same as that of the damage feature identification module and the micro-damage feature parameter extraction module in embodiment one, and will not be described in detail in this embodiment.

[0052] Step 3: Calculate the equivalent damage variable of the lithological sample based on the membership function of different microscopic damage characteristic parameters. ; Optionally, step three includes the following sub-steps: (1) The micro-damage characteristic parameters are mapped one-to-one with the damage variable D to form a feature / damage statistics table as the basis for identifying the mapping pattern, and the segment threshold of each micro-damage characteristic parameter is determined according to the feature / damage statistics table. (2) Based on the segmented thresholds of the micro-damage characteristic parameters, construct the membership functions of different micro-damage characteristic parameters; The SEM-damage mapping principle of membership function is as follows: taking the damage variable as the calibration target, the micro-damage characteristic parameters are converted into damage characterization quantities in the same dimension space through the membership function, and then the equivalent damage variable is obtained by fusing and mapping the values ​​of each membership function. This idea is consistent with the quantitative study of rock micro-damage "characterizing pre-failure damage by micro-crack / pore growth".

[0053] (3) Calculate the equivalent damage variables of lithological samples based on the membership functions of different micro-damage characteristic parameters. ; (4) Equivalent damage variables based on lithological samples The weighting coefficients corresponding to the micro-damage characteristic parameter S are optimized based on the damage variable D corresponding to the lithological sample. The objective function for optimizing the weighting coefficients corresponding to the microscopic damage characteristic parameter S is: ; In the formula, min represents the minimum value. It is an absolute value.

[0054] Example 3: Example 3 provides a rock damage quantification system based on SEM electron microscopy image recognition, used to implement the method described in Example 1. The system includes: The SEM image acquisition and preprocessing module is used to acquire SEM images of the rock sample under different magnification conditions and the corresponding imaging metadata, and to preprocess the SEM images based on the imaging metadata. The damage feature recognition module identifies damage features of the rock sample based on the preprocessed SEM electron microscope image of the rock sample to be tested. The micro-damage feature parameter extraction module extracts the micro-damage feature parameters of the rock sample to be tested based on the damage feature identification results. The rock damage quantification module calculates the membership function value of each micro-damage characteristic parameter based on the micro-damage characteristic parameters, and then calculates the equivalent damage variable of the rock sample to be tested, and classifies the damage level of the rock sample to be tested.

[0055] Example 4: Example 4 provides an electronic device, including at least one processor; and a memory communicatively connected to the processor; wherein the memory stores instructions to be executed by the processor, the instructions being executed by the processor to enable the processor to perform the method for quantifying rock damage in the rock sample to be tested as described in Example 1.

[0056] The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a GPU BOX, mobile phone, tablet computer, laptop computer, PDA, mobile Internet device (MID), robot, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), etc. The embodiments of this application do not specifically limit it.

[0057] The electronic device in this application embodiment can also be a device with an operating system. This operating system can be Android, Linux, Windows, or other possible operating systems; this application embodiment does not specifically limit the specific operating system used.

[0058] This application also provides a computer-readable storage medium storing a computer program / instruction thereon, which, when executed by a processor, implements the method for quantifying rock damage in a rock sample as described in Embodiment 1 of this application.

[0059] This application also provides a computer program product that, when run on an electronic device, enables the processor to execute the method for quantifying rock damage in the rock sample to be tested, as described in Embodiment 1 of this application.

[0060] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for quantifying rock damage based on SEM (Semiconductor Electron Microscopy) image recognition, characterized in that, The method includes: SEM images of the rock sample under different magnification conditions and corresponding imaging metadata were acquired, and the SEM images were preprocessed based on the imaging metadata. Damage features of the rock sample were identified based on the preprocessed SEM images of the rock sample. Based on the damage feature identification results, the microscopic damage feature parameters of the rock sample to be tested are extracted. Calculate the membership function value for each micro-damage characteristic parameter. ; Based on the membership function values ​​of different micro-damage characteristic parameters The equivalent damage variables of the rock sample under test are calculated by taking the weighting coefficients corresponding to different micro-damage characteristic parameters and classifying the damage level of the rock sample under test.

2. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 1, characterized in that, The microscopic damage characteristics of the rock sample to be tested include: fracture length density L d , percentage of fracture area A c C, the parameter of fracture connectivity n Porosity and texture entropy T e .

3. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 2, characterized in that, Membership function values ​​of micro-damage characteristic parameters The calculation method is as follows: ； In the formula, x represents the calculated value of the microscopic damage characteristic parameter, a S and b S Let S be the two segmented thresholds for the microscopic damage characteristic parameter S, and , This represents the membership function value corresponding to the calculated value x of the micro-damage characteristic parameter S; the micro-damage characteristic parameter S includes L d A c C n , and T e .

4. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 3, characterized in that, Equivalent damage variable of the rock sample to be tested The calculation method is as follows: ； In the formula, Let be the weighting coefficients corresponding to the microscopic damage feature parameter S, and n represent the number of microscopic damage feature parameters S. When S = 1, 2, 3, 4, 5, they correspond to L respectively. d A c C n , and T e .

5. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 4, characterized in that, The method for determining the weighting coefficients corresponding to the microscopic damage characteristic parameter S is as follows: SEM images of lithological samples corresponding to the rock sample under test at different damage evolution stages were obtained and preprocessed. At the same time, the damage variable D corresponding to the lithological samples at different damage evolution stages was obtained. Based on the preprocessed SEM electron microscope images of the lithological samples, damage features of the lithological samples are identified, and microscopic damage feature parameters of the lithological samples are extracted. Based on the membership functions of different microscopic damage characteristic parameters, the equivalent damage variables of lithological samples are calculated. ; Equivalent damage variables based on lithological samples The weighting coefficients corresponding to the micro-damage characteristic parameter S are optimized based on the damage variable D corresponding to the lithological sample.

6. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 5, characterized in that, The method for calculating the damage variable D corresponding to the lithological sample is as follows: ； In the formula, E0 is the elastic modulus in the undamaged state, and E is the elastic modulus of the lithological sample at the damage evolution stage.

7. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 6, characterized in that, The objective function for optimizing the weighting coefficients corresponding to the micro-damage characteristic parameter S is: ； In the formula, min represents the minimum value. It is an absolute value.

8. The rock damage quantification method based on SEM electron microscopy image recognition according to claim 1, characterized in that, Preprocessing of SEM electron microscope images based on imaging metadata includes: Scale calibration and unification of SEM electron microscope images based on imaging metadata; Gray-level normalization and noise suppression were performed on the scale-calibrated and unified SEM electron microscope images; Local contrast enhancement and feature saliency enhancement were performed on the noise-suppressed SEM images; The SEM electron microscope images after local contrast enhancement and feature saliency processing are cropped and scaled for output.

9. A rock damage quantification system based on SEM electron microscopy image recognition, used to implement the method according to any one of claims 1-8, characterized in that, The system includes: The SEM image acquisition and preprocessing module is used to acquire SEM images of the rock sample under different magnification conditions and the corresponding imaging metadata, and to preprocess the SEM images based on the imaging metadata. The damage feature recognition module identifies damage features of the rock sample based on the preprocessed SEM electron microscope image of the rock sample to be tested. The micro-damage feature parameter extraction module extracts the micro-damage feature parameters of the rock sample to be tested based on the damage feature identification results. The rock damage quantification module calculates the membership function value of each micro-damage characteristic parameter based on the micro-damage characteristic parameters, and then calculates the equivalent damage variable of the rock sample to be tested, and classifies the damage level of the rock sample to be tested.

10. An electronic device, characterized in that: The device includes at least one processor; and a memory communicatively connected to the processor; wherein the memory stores instructions executable by the processor to enable the processor to perform the method according to any one of claims 1-8.

Images