Identification and calculation method, device and readable storage medium for pore characteristics of porous materials

Abstract

The present invention relates to the field of feature measurement of porous materials, and particularly to a method, device and readable storage medium for identifying and calculating pore characteristics of porous materials. The method includes: obtaining a grayscale image of a porous material; performing a binarization operation on the grayscale image by using a multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic to obtain a binary image; using a multi-stage morphological connectivity refinement segmentation method to perform refined identification and segmentation on the binary image and calculate a first pore characteristic parameter; using a morphological-based Morse-Smale complex topological feature segmentation method to perform refined identification and segmentation on the binary image and calculate a second pore characteristic parameter; calculating the percentage difference between the first and second pore characteristic parameters; if the absolute value of the percentage difference is greater than a preset value, recalculate the first and second pore characteristic parameters; otherwise, record the average value as the final result. The present invention significantly improves the recognition accuracy and adaptability of complex pore images.

Description

Technical Field

[0001] The present invention relates to the field of feature measurement of porous materials, and particularly to a method, device and readable storage medium for identifying and calculating the pore characteristics of porous materials. Background Art

[0002] Porous materials have many desirable physical properties and can be widely used in the medical field, chemical engineering, civil engineering, etc. The pore characteristics are key parameters characterizing the performance of porous materials and can directly affect the application effects of materials in fields such as filtration, catalysis, energy storage, etc.

[0003] Traditional pore measurement methods mainly include gas adsorption method, mercury intrusion method, etc. These methods analyze the pore characteristics through physical experimental means and can provide overall parameters such as porosity. However, these methods generally have deficiencies such as long test cycles, complex experimental conditions, strong destructiveness to samples, etc. And due to the limitations of experimental conditions, the measurement results often lack accurate descriptions of the pore geometric characteristics and are difficult to meet the diverse needs of porous material characterization.

[0004] In recent years, with the rapid development of computer vision and artificial intelligence technologies, significant progress has been made in pore recognition using image recognition technology. Common recognition methods include threshold segmentation, edge detection, watershed algorithm, etc. These algorithms have the advantages of high computational efficiency, simple implementation and strong intuitiveness, and are particularly suitable for scenarios with high image quality, simple background and obvious pore characteristics. At the same time, these methods usually do not require a large amount of data support and have good adaptability to small-scale tasks and real-time processing requirements.

[0005] However, the currently commonly used image processing algorithms still face challenges in practical applications. On the one hand, there is a lot of noise and uneven contrast in the images of porous materials, which easily affects the segmentation accuracy between the pore and background regions. For example, the threshold segmentation method is sensitive to noise, difficult to process complex backgrounds, and the threshold selection is subjective with low robustness. On the other hand, a single algorithm often has difficulty in balancing recognition accuracy and robustness when processing images with complex pore morphologies, uneven feature distributions and uneven illumination, and for complex connected pores, the recognition and refinement effects of traditional methods (such as threshold segmentation, edge detection, watershed algorithm, etc.) are limited. In addition, there is a lack of a collaborative optimization mechanism between different algorithms, making them show certain instability when extracting multi-dimensional pore characteristics (such as porosity, geometric morphology, connectivity, etc.). Therefore, it is of great significance to develop a comprehensive method that combines the advantages of multiple algorithms and can achieve high-precision recognition through differential analysis and parameter optimization. Summary of the Invention

[0006] The features and advantages of the present invention are partially stated in the following description, or may be obvious from the description, or may be learned by practicing the present invention.

[0007] To overcome the problems of the prior art, the present invention provides a method for identifying and calculating pore characteristics of a porous material, comprising the steps of:

[0008] Obtain a two-dimensional image of the porous material, and preprocess the two-dimensional image to obtain a grayscale image;

[0009] Perform a binarization operation on the grayscale image by using a multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic to obtain a binary image;

[0010] Use a multi-stage morphological connectivity refinement segmentation method to perform refined identification and segmentation on the binary image and calculate pore characteristic parameters, denoted as the first pore characteristic parameters;

[0011] Use a Morse-Smale complex topological feature segmentation method based on morphology to perform refined identification and segmentation on the binary image and calculate pore characteristic parameters, denoted as the second pore characteristic parameters;

[0012] Calculate the percentage difference between the first pore characteristic parameters and the second pore characteristic parameters;

[0013] If the absolute value of the percentage difference is greater than a preset value, optimize the algorithm characteristic parameters in the multi-stage morphological connectivity refinement segmentation method and the Morse-Smale complex topological feature segmentation method based on morphology, and recalculate the first pore characteristic parameters and the second pore characteristic parameters; otherwise, record the average value of the first pore characteristic parameters and the second pore characteristic parameters as the final result.

[0014] In one embodiment of the present invention, the step of performing a binarization operation on the grayscale image by using a multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic to obtain a binary image includes:

[0015] S21. Perform a global grayscale analysis on the grayscale image, and use a mean shift clustering method to determine an initial threshold;

[0016] S22. Divide the image into multiple local regions, and capture the characteristics of each local region in real time. The characteristics of the local region include contrast and texture characteristics;

[0017] S23. Dynamically adjust the local threshold according to the characteristics of the local region in combination with fuzzy logic rules;

[0018] S24. Fuse the initial threshold and the local threshold using a region growing weight model to generate a final fused threshold;

[0019] S25. Iteratively optimize the threshold through an expectation maximization algorithm to obtain a final optimized threshold;

[0020] S26. Binarize the grayscale image according to the final optimization threshold.

[0021] In one embodiment of the present invention, the use of the multi-stage morphological connectivity refinement segmentation method to perform refined identification and segmentation on the binarized image and calculate pore characteristic parameters includes:

[0022] S31. Perform preprocessing for pore extraction on the binarized image;

[0023] S32. Mark the connected regions formed by all foreground pixels according to the processed binarized image;

[0024] S33. Perform morphological reconstruction based on the marked connected regions;

[0025] S34. Refine the distinction between pores and the background for the image after morphological reconstruction.

[0026] In one embodiment of the present invention, step S34 includes: performing a distance transform on the image after morphological reconstruction, and performing watershed segmentation through the extreme points on the distance transform image to refine the boundary between connected pores and the background and accurately segment the connected regions; wherein, the calculation expression of the distance transform is as follows:

[0027]

[0028] In the expression represents the distance transform value of pixel ; represents the Euclidean distance between two pixels.

[0029] In one embodiment of the present invention, the use of the Morse-Smale complex topological feature segmentation method based on morphology to perform refined identification and segmentation on the binarized image and calculate pore characteristic parameters specifically includes:

[0030] S41. Generate a gradient field through the gradient calculation of the grayscale image for scalar field construction;

[0031] S42. Based on the preprocessed gradient field, construct a scalar field , and define key points;

[0032] S43. Construct Morse-Smale complex segmentation units through streamline tracing according to the key points to obtain pore regions.

[0033] In one embodiment of the present invention, the key points in step S42 include maximum points, minimum points, and saddle points; the classification of the key points is determined by the eigenvalues of the Hessian matrix 𝐻(𝑓):

[0034]

[0035] When it reaches the extreme point, it reaches the saddle point when

[0036] In one embodiment of the present invention, the Morse-Smale complex segmentation unit in step S43 is formed by the intersection of the upper-manifold and the lower-manifold; wherein, the upper-manifold is a set of pixels where the gradient ascends to the maximum point; the lower-manifold is a set of pixels where the gradient descends to the minimum point.

[0037] In one embodiment of the present invention, the preset value is not greater than 1.0%.

[0038] The present invention also provides a porous material pore feature recognition and calculation device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the porous material pore feature recognition and calculation method according to any one of the present invention.

[0039] The present invention provides a computer-readable storage medium, on which a porous material pore feature recognition and calculation program is stored. When the porous material pore feature recognition and calculation program is executed by a processor, it implements the steps of the porous material pore feature recognition and calculation method according to any one of the present invention.

[0040] The present invention provides a porous material pore feature recognition and calculation method, device, and computer-readable storage medium. Through the innovative combination of dynamic fuzzy logic, multi-stage morphological thinning segmentation method, and improved Morse-Smale complex segmentation algorithm, a high-precision and high-robustness pore feature recognition method is constructed.

[0041] The present invention overcomes the limitations of traditional physical tests and single image recognition algorithms. It can not only quickly and automatically extract the geometric and topological features of porous materials, but also significantly improve the recognition accuracy and adaptability of complex pore images, providing strong technical support for the performance research and application design of porous materials.

[0042] By reading the specification, those of ordinary skill in the art will better understand the features and contents of these technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The present invention will be specifically described below with reference to the drawings and examples. The advantages and implementation manners of the present invention will become more obvious. The content shown in the drawings is only for the explanation of the present invention and does not constitute any limitation to the present invention. In the drawings:

[0044] Figure 1Schematic flowchart of the method for identifying the pore characteristics of the porous material according to the embodiment of the present invention;

[0045] Figure 2 is the grayscale histogram of the grayscale image in this embodiment;

[0046] Figure 3 Schematic flowchart of the multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic;

[0047] Figure 4 Schematic flowchart of the multi-stage morphological connectivity refinement segmentation method;

[0048] Figure 5 Schematic flowchart of the morphological-based Morse-Smale complex topological feature segmentation method;

[0049] Figure 6 is the comparison diagram before and after image processing in this embodiment. Detailed implementation manners

[0050] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0051] Embodiment 1

[0052] As Figure 1 shown, this embodiment provides a method for identifying and calculating the pore characteristics of a porous material, including the steps of:

[0053] S10. Obtain a two-dimensional image of the porous material, and preprocess the two-dimensional image to obtain a grayscale image;

[0054] Generally, MATLAB (MATrix LABoratary) can be used to preprocess the two-dimensional image;

[0055] S20. Perform a binarization operation on the grayscale image by using a multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic to obtain a binarized image;

[0056] Through the binarization operation, the grayscale image can be converted into a black-and-white image, and the pixels are classified into "foreground" (usually the pore region) and "background" (usually the solid region) through a threshold. In the present invention, the pore region includes independent pore regions and connected regions.

[0057] S30. Use the multi-stage morphological connectivity refinement segmentation method to perform refined identification and segmentation on the binarized image and calculate the pore characteristic parameters, denoted as the first pore characteristic parameters;

[0058] S40. Use the Morse-Smale complex topological feature segmentation method based on morphology to refine the recognition and segmentation of the binary image and calculate the pore feature parameters, denoted as the second pore feature parameters;

[0059] It should be noted that steps S30 and S40 have no sequence relationship, and the two can also be carried out simultaneously. The processing scope of steps S30 and S40 is the entire binary image, but the ultimate goal of these two methods is to identify and segment the pore region. Among them, the multi-stage morphological connectivity refinement segmentation method in step S30 focuses on geometric feature extraction, while the Morse-Smale complex segmentation algorithm in step S40 focuses on topological structure analysis.

[0060] S50. Calculate the percentage difference between the first pore feature parameters and the second pore feature parameters; determine whether the absolute value of the percentage difference is greater than a preset value. If so, optimize the algorithm feature parameters in steps S30 and S40 and return to step S30; if not, proceed to step S60;

[0061] In specific implementation, the algorithm feature parameters can be reversely optimized according to the calculation results of the two segmentation algorithms, and the difference can be continuously reduced through iteration. By automatically adjusting the parameters of the multi-stage morphological connectivity refinement segmentation method and the improved Morse-Smale complex topological feature segmentation algorithm based on morphology, re-perform S30 and S40 until the absolute value of the percentage difference is less than the preset value. Generally, the preset value is not greater than 1.0%, for example, it is 1.0%.

[0062] The optimization of the above algorithm feature parameters can be achieved by adjusting common operations in the fields of computer vision and image processing such as the size of the structural element in morphological operations and the gradient threshold in topological analysis. Among them, the algorithm feature parameters that can be changed in step S30 include at least one of the shape and size of the structural element (such as circular, rectangular and their dimensions), the determination rule of the connected region (such as 8-neighborhood or 4-neighborhood), and the iteration times of opening operation and closing operation. The algorithm feature parameters that can be changed in step S40 include at least one of the gradient threshold (affecting the accuracy of streamline tracking), the construction rule of the scalar field (such as smoothness, the threshold for defining key points), and the segmentation scale of the upper-manifold and the lower-manifold.

[0063] The present invention combines two segmentation algorithms (morphological connectivity refinement segmentation method and Morse-Smale complex topological segmentation method), and iteratively optimizes the feature parameters through differential analysis. This optimization is better than the single optimization result and has significant technical effects (improving the recognition accuracy of complex pores).

[0064] S60. Denote the average value of the first pore feature parameters and the second pore feature parameters as the final result.

[0065] In this embodiment, a pore feature recognition system with accuracy, efficiency and robustness is constructed through the global robustness of dynamic fuzzy logic, the geometric refinement segmentation capability of multi-stage morphology and the complex structure analysis capability of the Morse-Smale topological algorithm. The collaborative work overcomes the shortcomings of a single algorithm and shows significant advantages in adapting to complex image characteristics, accurately identifying complex pore structures and comprehensively extracting multi-dimensional features.

[0066] Embodiment 2

[0067] On the basis of the above-mentioned first embodiment, step S10 in the method for identifying and calculating pore characteristics of porous materials provided in this embodiment specifically includes the following steps:

[0068] S11, collecting a two-dimensional image of the porous material by a digital image processing device, and recording it as an image to be processed;

[0069] The two-dimensional image is collected by digital image processing equipment such as SEM, CT or industrial high-precision camera, preferably with dpi ≥ 300 and image size ≥ 1000 × 1000 pixels. The image format supports common formats such as bmp, png, jpg, tiff, etc., and tiff format is preferred.

[0070] S12, convert the two-dimensional image into a grayscale image with a grayscale value range of [0,255], and display it as follows Figure 2 Grayscale histogram of the grayscale image shown.

[0071] In specific implementation, the original image can also be cut by the coordinate value of the BoundingBox before step S12 to obtain the interested region of the image to be processed. At this time, it is only necessary to convert the interested region of the image to be processed into a grayscale image.

[0072] S13. Select different image enhancement methods according to the usage scenario to enhance the image and obtain the final grayscale image.

[0073] Among them, the image enhancement methods can be adaptive histogram equalization (suitable for pore images with poor contrast), Gaussian filtering (suitable for pore images with more noise), bilateral filtering (suitable for images that need to remove noise but retain pore boundaries), and unsupervised image segmentation (suitable for images with certain differences between pores and background).

[0074] Embodiment 3

[0075] Based on the above embodiment 1, Figure 3 As shown, step S20 in the method for identifying and calculating pore characteristics of porous materials provided in this embodiment specifically includes the following steps:

[0076] S21. Dynamic threshold initialization: Perform global grayscale analysis on the input grayscale image, and use the mean shift clustering method to determine the initial threshold , enhancing the robustness to images with uneven brightness. According to the characteristics of the image grayscale histogram, the mean shift clustering calculation expression is as follows:

[0077]

[0078] In the expression is the grayscale value of the image pixel; is the clustering center; is the mean shift bandwidth parameter, which is used to control the size of the clustering window.

[0079] S22. Real-time feature monitoring: Divide the grayscale image into multiple local regions, and capture the contrast and texture features of each local region in real time as the input basis for subsequent dynamic threshold adjustment. The calculation expressions for local contrast and local texture complexity are as follows:

[0080]

[0081]

[0082] In the expression is the local contrast; and are the maximum and minimum grayscale values of the local region respectively; is the local texture complexity; and are the second-order derivatives of the image in the x and y directions; is the number of pixels in the local region.

[0083] S23. Local threshold adjustment: According to the characteristics of the local region monitored in real time, dynamically adjust the local threshold through fuzzy logic rules, avoiding the limitations of fixed rules and ensuring adaptation to the characteristics of different image regions. Adjusting the local threshold based on the fuzzy logic method is achieved through the following formula:

[0084]

[0085] In the expression, 𝑘 is the fuzzy function adjustment coefficient.

[0086] S24. Global and local threshold fusion: Combine the global initial threshold and the local dynamic threshold, and use the weight model of region growing for fusion to generate the final fused threshold , and the calculation expression is as follows:

[0087]

[0088] In the expression is the pixel gradient magnitude, is the saliency mask (the salient region is 1 and the rest is 0).

[0089] S25. Iterative optimization: The threshold is iteratively optimized through the expectation-maximization algorithm to further improve the segmentation accuracy. The expressions for expectation and maximization calculations are as follows:

[0090]

[0091]

[0092] and represent the probabilities of the foreground and background respectively; is the Gaussian distribution; and are the means of the foreground and background respectively. When the threshold changes between two adjacent iterations or reaches the maximum number of iterations, the iteration terminates. The maximum number of iterations is recommended to be 10 - 15 times, preferably 12 times, is recommended to be taken as 0.5.

[0093] S26. Final binarization: According to the finally optimized threshold the image is binarized, and the binarization calculation expression is as follows:

[0094]

[0095] In the expression is the grayscale image; is the binary image; represents the grayscale threshold.

[0096] This embodiment adopts a multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic, which can dynamically adjust the threshold according to the contrast and texture characteristics of the local region of the image, and achieve precise binarization through global and local fusion. Compared with the traditional fixed-threshold segmentation method, this method has higher robustness and segmentation accuracy for images with uneven illumination, complex gray distribution, and background noise, effectively avoiding the problems of false detection or missed detection.

[0097] Embodiment 4

[0098] Based on any of the above embodiments, as Figure 4 shown, the steps of step S30 in the method for identifying and calculating the pore characteristics of the porous material provided in this embodiment specifically include the steps:

[0099] S31. Perform preprocessing for pore extraction on the binarized image;

[0100] Based on the preliminary binarization, first use the opening operation to eliminate small noises, and then use the closing operation to fill the small holes inside the pores. The calculation expressions for the opening operation and the closing operation are as follows:

[0101]

[0102]

[0103] In the expressions A is the input image; 𝐵 is the structuring element, which is circular with a radius of 1 / 3 of the average pore diameter; represents the erosion operation; represents the dilation operation.

[0104] S32. Label the connected regions formed by all foreground pixels based on the processed binarized image;

[0105] Perform connected component analysis on the processed binarized image, select 8-neighborhood connectivity to detect all connected parts, label each connected region, and determine whether it belongs to a pore. The calculation expression is as follows:

[0106]

[0107] In the expressions is the label value at position in the image; is the label of a certain connected region.

[0108] The connected regions in the present invention are preliminary labels for all foreground pixels (i.e., pore regions) in the processed binarized image. These regions include at least one of the independent regions of pores (single pores) and the connected regions of pores (regions formed by multiple interconnected pores). Among them, the connected region of pores is a special case of the connected region, referring to a region composed of multiple interconnected pores. From the definition, the connected region of pores is a part of the connected region, but the connected region may also contain other noises or non-pore features (such as possible misjudged regions).

[0109] S33. Perform morphological reconstruction based on the labeled connected regions to obtain the complete pore region;

[0110] The specific calculation expression is as follows:

[0111]

[0112] Among them, is the image to be reconstructed (i.e., the preliminarily labeled connected region); is the structuring element, which is circular with a radius of 1 / 3 of the average pore diameter; It is the result of morphological reconstruction.

[0113] S34. Refine the distinction between pores and the background for the image after morphological reconstruction;

[0114] In this embodiment, the watershed algorithm is used to refine the distinction between pores and the background. More specifically, first perform a distance transform on the image after the morphological reconstruction, and then perform watershed segmentation through the extreme points on the distance transform image to refine the boundary between connected pores and the background and accurately segment the connected regions;

[0115] Among them, the calculation expression of the distance transform is as follows:

[0116]

[0117] In the expression represents the pixel 's distance transform value; represents the Euclidean distance between two pixels.

[0118] It can be seen that the connected regions marked in step S32 (including independent pores and connected regions of pores) will be further refined and distinguished in subsequent steps (such as morphological reconstruction or watershed segmentation), and the connected regions of pores can be accurately extracted from them.

[0119] S35: Calculate the pore geometric feature parameters and denote them as the first pore geometric feature parameters.

[0120] The calculation of the pore geometric feature parameters can be based on the refined binary image.

[0121] In this embodiment, step S30 focuses on geometric feature extraction and uses a multi-stage morphological method to finely identify and segment all pores in the entire binary image, including the refinement of pore boundaries, connectivity detection, etc.

[0122] Embodiment Five

[0123] Based on any of the above embodiments, as Figure 5 shown, step S40 in the method for identifying and calculating the pore characteristics of the porous material provided in this embodiment specifically includes the steps:

[0124] S41: Generate a gradient field through the gradient calculation of the grayscale image for scalar field construction;

[0125] The formula for scalar field construction is as follows:

[0126]

[0127] and is the image at x andy The gradient component in the [direction].

[0128] Generally, before generating the gradient field, preprocessing of pore extraction can be performed first. That is, on the basis of preliminary binarization, the opening operation is first used to eliminate small noises, and then the closing operation is used to fill the small holes inside the pores. The calculation expressions of the opening operation and the closing operation are as follows:

[0129]

[0130]

[0131] In the expression A is the input image; 𝐵 is the structuring element, which is circular and the radius is taken as 1 / 3 of the average diameter of the pores; represents the erosion operation; represents the dilation operation.

[0132] S42. Based on the preprocessed gradient field, construct a scalar field , and define key points;

[0133] In this embodiment, the key points include maximum points, minimum points, and saddle points; among them, the maximum point (maximum value) is the position where the gradient is zero and is the maximum value in the local area; the minimum point (minimum value) is the position where the gradient is zero and is the minimum value in the local area; the saddle point is the position where the gradient is zero but is neither the maximum value nor the minimum value.

[0134] The classification of key points is determined by the eigenvalues of the Hessian matrix 𝐻(𝑓):

[0135]

[0136] When take the extreme points, take the saddle points.

[0137] S43: Construct Morse-Smale complex segmentation units through streamline tracing according to the key points to obtain the pore region.

[0138] It should be noted that the segmentation units in this step are based on the topological features of the grayscale image and have no direct association with the connected regions in the binary image.

[0139] When performing streamline tracing, starting from each pixel in the image, trace along the gradient descent direction to the nearest minimum point to form an attracting streamline; trace along the gradient ascent direction to the nearest maximum point to form a repulsive streamline:

[0140] (Attracting streamline)

[0141] (Repulsive streamline)

[0142] In the expression is the current pixel coordinate, is the streamline step size.

[0143] The Morse-Smale complex segmentation unit constructs the upper-manifold and the lower-manifold through attractive streamlines and repulsive streamlines, and generates them through the intersection of the two. Specifically, the upper-manifold is the set of pixels where the gradient ascends to the maximum point; the lower-manifold is the set of pixels where the gradient descends to the minimum point. The intersection of the upper-manifold and the lower-manifold can form the Morse-Smale complex segmentation unit. The construction of the Morse-Smale complex segmentation unit can refine the segmented pore region and ensure accurate topological structure analysis.

[0144]

[0145] Where and are the upper-manifold and the lower-manifold respectively.

[0146] S44: Extract the pore geometric features from the pore region, denoted as the second pore feature parameter.

[0147] The Morse-Smale complex segmentation algorithm in this embodiment focuses on topological structure analysis (such as gradient field, scalar field, key points, and streamline tracing), and is a global topological analysis method that can further identify complex pore topologies (such as connectivity and overall distribution), while the multi-stage morphological connectivity refinement segmentation method focuses on geometric feature extraction. The combination of geometric and topological features can take into account the connectivity, boundary fineness, and morphological complexity of pores simultaneously, thus significantly improving the ability to identify complex pores.

[0148] Example Six

[0149] In this embodiment, the pore feature recognition calculation method provided in any of the above embodiments is used to calculate the surface porosity of a specimen, and the results are shown in Table 1 below. The images before and after processing are as Figure 6 shown.

[0150] Table 1 Recognition calculation results of concrete surface pore image features

[0151]

[0152] The conclusion obtained by measuring the porosity of the same surface of the specimen using the traditional physical experiment gas adsorption method is 8.19%.

[0153] The above traditional physical experiment gas adsorption method is specifically implemented through the following steps:

[0154] S91. Seal the remaining five surfaces of the test piece of the embodiment, only expose the surface to be measured, and measure the total area of the exposed surface. ;

[0155] S92. Dry the test piece at 100 - 105 °C to remove the residual moisture in the pores and prevent it from affecting gas adsorption.

[0156] S93. Use nitrogen as the adsorption gas, put the dried test piece into the sample chamber of the gas adsorption instrument, and ensure that the pores of the exposed surface are in direct contact with the gas:

[0157] S94. Adjust the gas pressure and record the gas adsorption amount at different relative pressures ( , that is, the ratio of the actual pressure to the saturated vapor pressure).

[0158] S95. Measure and record the isothermal adsorption data (the relationship between the gas adsorption amount and ) after adsorption equilibrium.

[0159] S96. Use the BET equation to fit the isothermal adsorption curve and calculate the specific surface area :

[0160]

[0161] In the expression is the monolayer adsorption amount; is Avogadro's constant ( ); is the effective area occupied by the adsorption gas molecules, and for nitrogen it is ; is the mass of the sample.

[0162] S97. Assume that the adsorption gas only acts on the pore region, then there is . According to the total exposed area and the pore area , the porosity is:

[0163]

[0164] Through comparative test results, the porosity measured by the method of the present invention is 8.37%. Compared with 8.19% measured by the traditional gas adsorption method, the error is only 2.2%, indicating that the two have high consistency in measurement accuracy. This tiny error mainly stems from the differences in pore structure definition and test conditions between the two methods. The present invention is based on image recognition and computational algorithms, and uses digital means to directly characterize pore characteristics, avoiding the uncertainties caused by experimental environment and gas molecule behavior in traditional physical methods, showing high reliability and stability. Therefore, this method is applicable to the pore measurement of porous materials and has more flexible application potential.

[0165] Compared with traditional physical test methods, the present invention can efficiently extract all-round data including various geometric features such as pore area, perimeter, length, width, curvature, and porosity through only one image. This not only significantly improves the dimension of pore feature analysis but also avoids the limitation that traditional methods can only obtain limited data in one experiment. In addition, traditional methods often require complex equipment operation and a long test cycle, while the present invention greatly simplifies the experimental process through automated processing and digital analysis, greatly improving the test efficiency and data utilization rate, and is particularly suitable for the rapid analysis of a large number of samples.

[0166] Example Seven

[0167] Based on the above embodiments, the present invention also proposes a device for a method of identifying and calculating pore characteristics of a porous material. The device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the method of identifying and calculating pore characteristics of a porous material as described in any one of the above.

[0168] It should be noted that the above device embodiment and method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment, and the technical features in the method embodiment are correspondingly applicable to the device embodiment, and will not be elaborated here.

[0169] Example Eight

[0170] Based on the above embodiments, the present invention also proposes a computer-readable storage medium. A program for a method of identifying and calculating pore characteristics of a porous material is stored on the computer-readable storage medium. When the program for a method of identifying and calculating pore characteristics of a porous material is executed by a processor, it implements the steps of the method of identifying and calculating pore characteristics of a porous material as described in any one of the above.

[0171] It should be noted that the above medium embodiment and method embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment, and the technical features in the method embodiment are correspondingly applicable to the medium embodiment, and will not be elaborated here.

[0172] The pore feature recognition calculation method, device and computer-readable storage medium provided by the present invention achieve multi-scale adaptive binarization of images in complex scenarios by means of a multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic, overcoming the deficiencies of the threshold segmentation method; according to the pore feature parameters calculated by the multi-stage morphological connectivity refinement segmentation method and the improved Morse-Smale complex segmentation algorithm based on morphology, the algorithm parameters are reversely optimized through differential analysis. The optimization process makes the algorithm adapt to diverse pore structures and image characteristics through multiple rounds of iteration, improving the adaptability and robustness to complex, multi-scale, and diverse pores. It overcomes the problems that traditional algorithms (such as simple threshold segmentation and single morphological methods) often cannot take into account both geometric and topological features simultaneously when dealing with complex or morphologically variable pores, and are prone to problems such as missed detection, false detection, blurred boundaries, poor adaptability, and insufficient recognition ability for complex structures, providing strong technical support for the efficient characterization of pore features of porous materials.

[0173] It should be noted that in this article, the terms "including", "comprising" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including that element.

[0174] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0175] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disc) and includes several instructions for causing a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0176] The embodiments of the present invention have been described above in conjunction with the accompanying drawings, but the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative and not restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit and scope protected by the claims of the present invention, and these all belong to the protection scope of the present invention.

Claims

1. A method for calculating pore characteristics of porous materials, characterized in that: Includes steps: Acquire a two-dimensional image of the porous material, and preprocess the two-dimensional image to obtain a grayscale image; A multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic is used to perform a binarization operation on the grayscale image to obtain a binarized image; Using a multi-stage morphological connected refinement segmentation method, the binary image is refinedly identified and segmented, and pore characteristic parameters are calculated, which are recorded as first pore characteristic parameters; Using a morphologically based Morse-Smale complex topological feature segmentation method, the binary image is refinedly identified and segmented, and pore characteristic parameters are calculated, which are recorded as second pore characteristic parameters; Calculating a percentage difference between the first pore characteristic parameter and the second pore characteristic parameter; If the absolute value of the percentage difference is greater than a preset value, the algorithm characteristic parameters in the multi-stage morphological connectivity refinement segmentation method and the morphology-based Morse-Smale complex topological feature segmentation method are optimized, and the first pore characteristic parameter and the second pore characteristic parameter are recalculated; otherwise, the average value of the first pore characteristic parameter and the second pore characteristic parameter is recorded as the final result.

2. The porous material pore feature recognition and calculation method according to claim 1, characterized in that: The multi-scale hybrid threshold segmentation method driven by dynamic fuzzy logic is used to perform a binarization operation on the grayscale image to obtain a binarized image, which includes: S21, performing global grayscale analysis on the grayscale image, and determining an initial threshold using a mean shift clustering method; S22, dividing the image into multiple local areas, and capturing characteristics of each local area in real time, wherein the local area characteristics include contrast and texture characteristics; S23, dynamically adjusting the local threshold according to the local area characteristics combined with fuzzy logic rules; S24, fusing the initial threshold and the local threshold using a region growing weight model to generate a fused final threshold; S25, iteratively optimizing the threshold value through the expectation maximization algorithm to obtain the final optimized threshold value; S26. Binarize the grayscale image according to the final optimized threshold.

3. The porous material pore feature recognition and calculation method according to claim 1, characterized in that: The method of using the multi-stage morphological connectivity refinement segmentation method to perform fine identification and segmentation on the binary image and calculate pore characteristic parameters includes: S31, preprocessing the binary image for pore extraction; S32, marking the connected areas formed by all foreground pixels according to the processed binary image; S33, performing morphological reconstruction based on the marked connected regions; S34. Refine and distinguish the pores and the background of the image after morphological reconstruction.

4. The porous material pore feature recognition and calculation method according to claim 3, characterized in that: The step S34 includes: performing distance transformation on the morphologically reconstructed image, and performing watershed segmentation through the extreme points on the distance transformation image, so as to refine the boundary between the connected pores and the background and accurately segment the connected area; wherein the calculation expression of the distance transformation is as follows: In the expression Represents pixels The distance transformation value of Represents the Euclidean distance between two pixels.

5. The porous material pore feature recognition and calculation method according to claim 1, characterized in that: The method of using the morphological-based Morse-Smale complex topological feature segmentation method to perform fine identification and segmentation on the binary image and calculate pore feature parameters specifically includes: S41, generating a gradient field by gradient calculation of the grayscale image for scalar field construction; S42. Construct a scalar field based on the preprocessed gradient field , define key points; S43, constructing a Morse-Smale complex segmentation unit according to the key points through streamline tracing to obtain a pore area.

6. The porous material pore feature recognition and calculation method according to claim 5, characterized in that: The key points in step S42 include maximum points, minimum points and saddle points; the classification of the key points is determined by the eigenvalues ​​of the Hessian matrix 𝐻(𝑓): when When taking the extreme point, Take the saddle point.

7. The porous material pore feature recognition and calculation method according to claim 6, characterized in that: The Morse-Smale complex segmentation unit in step S43 is formed by the intersection of an upper-manifold and a lower-manifold; wherein the upper-manifold is a set of pixels whose gradient rises to a maximum point; and the lower-manifold is a set of pixels whose gradient falls to a minimum point.

8. The porous material pore feature recognition and calculation method according to claim 1, characterized in that: The preset value is no greater than 1.0%.

9. A porous material pore feature recognition computing device, characterized in that: The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the porous material pore feature identification calculation method as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a porous material pore feature recognition calculation program, which, when executed by a processor, implements the steps of the porous material pore feature recognition calculation method according to any one of claims 1 to 8.

Images