Quantitative evaluation method for fiber distribution characteristics of soil shear surface based on image recognition

Abstract

This invention relates to the fields of geotechnical engineering and image processing technology, specifically to a method for quantitatively evaluating the fiber distribution characteristics of soil shear surfaces based on image recognition. In this invention, a fiber distribution characteristic evaluation model based on morphological parameters is constructed to achieve a quantitative assessment of fiber distribution characteristics, providing fundamental data and theoretical support for the accurate description of the stress performance of fibrous soils and the establishment of their constitutive models. Specifically, fractal dimension and Weibull distribution characteristic parameters are introduced to achieve a comprehensive characterization of the fiber distribution morphology of soil shear surfaces, enabling quantitative evaluation and improving the rationality and accuracy of the assessment.

Description

Technical Field

[0001] This invention relates to the fields of geotechnical engineering and image processing technology, specifically to a method for quantitatively evaluating the fiber distribution characteristics of soil shear surfaces based on image recognition. Background Technology

[0002] Fiber materials are widely used in geotechnical engineering, composite materials, and biomedical materials. For example, in geotechnical engineering, fiber-reinforced soil is used to fill roadbeds, improving their bearing capacity, reducing settlement, and suppressing uneven deformation. The tensile strength of natural or synthetic fibers facilitates load transfer and redistribution, thereby reducing the risk of sudden brittle failure of the soil and significantly improving the engineering performance of fiber-reinforced soil. However, during the mixing of fibers and soil, localized fiber concentration or large-area agglomeration is prone to occur. This phenomenon has been clearly observed at both macroscopic and microscopic scales. This uneven distribution of fibers leads to the deterioration of the mechanical properties of fiber-reinforced soil composites.

[0003] Currently, fiber-reinforced soil is typically prepared using a mixing method. However, during the mixing process, it is impossible to precisely control the uniform dispersion of randomly distributed fibers throughout the sample. Therefore, characterizing the random distribution of fibers and analyzing their impact on mechanical behavior is particularly important. However, to date, quantitative characterization of fiber morphology within the soil matrix still suffers from several shortcomings. For example, traditional methods only perform simple area statistics, failing to comprehensively characterize the spatial complexity of fiber distribution; there is a lack of standardized modeling for fiber length distribution, making it impossible to determine the fiber length distribution characteristics at specific interfaces; and the observation area lacks uniform constraints, leading to poor repeatability due to arbitrary selection. Summary of the Invention

[0004] Based on this, the present invention provides a method for quantitative evaluation of fiber distribution characteristics of soil shear surface based on image recognition, which solves at least one problem in the prior art.

[0005] In a first aspect, the present invention provides a method for quantitatively evaluating the fiber distribution characteristics of soil shear surfaces based on image recognition, comprising the following steps: Acquire an image of a soil shear plane, wherein the soil contains fibers; The fibers in the image are identified based on the dual threshold method, and the fractal dimension of the fiber distribution is calculated based on the box method. The length of the fibers in the image is calculated based on the minimum bounding rectangle method, and the Weibull distribution characteristic parameters of the fiber length are statistically analyzed. Based on the fractal dimension and Weibull distribution characteristic parameters, a fiber distribution characteristic evaluation model is established using the weight allocation method. The aggregation index is calculated through the fiber distribution characteristic evaluation model, and the fiber distribution characteristics of the soil shear surface are determined based on the aggregation index.

[0006] In the present invention, a quantitative evaluation of the fiber distribution characteristics is achieved by constructing an evaluation model for fiber distribution characteristics based on morphological parameters, providing basic data and theoretical support for the accurate description of the mechanical properties of fiber-containing soil masses and the establishment of their constitutive models. Among them, the fractal dimension and Weibull distribution characteristic parameters are introduced to comprehensively characterize the fiber distribution morphology on the soil shear surface, achieving quantitative evaluation and improving the rationality and accuracy of the evaluation.

[0007] In some optional embodiments, the soil mass is fiber-reinforced soil prepared from red clay and bamboo fibers by the mixing method.

[0008] In some optional embodiments, the Weibull distribution characteristic parameters for statistically analyzing the fiber length include determining the shape parameter and scale parameter of the Weibull distribution according to the following probability density function; ; Where L represents the fiber length, k represents the shape parameter of the Weibull distribution, λ represents the scale parameter of the Weibull distribution.

[0009] In some optional embodiments, the calculation formula for the aggregation index is: ; Where AI represents the aggregation index, , and respectively represent the normalized fractal dimension, the shape parameter of the Weibull distribution, and the scale parameter of the Weibull distribution.

[0010] In some optional embodiments, the rules for determining the fiber distribution characteristics on the soil shear surface based on the aggregation index include: If the aggregation index is less than 0.45 (AI < 0.4), it is determined that the fiber distribution characteristics on the soil shear surface are well-dispersed; If the aggregation index is between 0.45 and 0.65 (0.45 < AI ≤ 0.65), it is determined that the fiber distribution characteristics on the soil shear surface are slightly aggregated; If the aggregation index is between 0.65 and 0.85 (0.65 < AI ≤ 0.85), it is determined that the fiber distribution characteristics on the soil shear surface are moderately aggregated; If the aggregation index is greater than 0.85, it is determined that the fiber distribution characteristics on the soil shear surface are severely aggregated.

[0011] In some optional embodiments, calculating the fractal dimension of the fiber distribution based on the box method includes: Define box size sequence ,in Indicates the first i The side length of the box, , Indicates the width of the image. Indicates the length of the image. This indicates taking the maximum value. Represents positive integers; For each box size The image is divided into For a given grid size, count the number of fractal boxes that contain at least one fiber pixel. ; Regarding box size and the corresponding number of boxes Take the natural logarithm to obtain the data. The least squares method is used to analyze the data. Perform a linear fit, and the linear fit equation is: ; in, The fractal dimension represents the fiber distribution in the image. This represents the fitted parameters.

[0012] In some optional embodiments, identifying fibers in the image based on a dual-threshold method includes: The image is preprocessed to grayscale to obtain a grayscale image; In the grayscale image, pixels with grayscale values ​​greater than a first preset threshold are set to 255, and pixels with grayscale values ​​less than or equal to the first preset threshold are set to 0, thereby generating a first intermediate image. In the grayscale image, pixels with grayscale values ​​greater than the second preset threshold are set to 0, and pixels with grayscale values ​​less than or equal to the second preset threshold are set to 255, thereby generating a second intermediate image. Perform a pixel-by-pixel logical AND operation on the first intermediate image and the second intermediate image to obtain the fiber mask image; The fibers in the image are determined based on the fiber mask image.

[0013] In some optional embodiments, the first preset threshold is the minimum gray value of the fiber, and the second preset threshold is the maximum gray value of the fiber.

[0014] In a second aspect, the present invention provides an electronic device comprising: At least one processor; and memory that is communicatively connected to at least one processor; The memory stores instructions that, when executed by at least one processor, implement the image recognition-based quantitative assessment method for fiber distribution characteristics of soil shear surfaces.

[0015] Thirdly, the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the image recognition-based quantitative evaluation method for fiber distribution characteristics of soil shear surfaces.

[0016] Due to the adoption of the above technical solutions, the embodiments of the present invention have at least the following beneficial effects: (1) Fractal dimension and Weibull distribution characteristic parameters were introduced to achieve a comprehensive characterization of the fiber distribution morphology on the soil shear surface. Among them, the fractal dimension characterizes the complexity and overall uniformity of the fiber-filled shear surface, and the Weibull distribution characteristic parameters ( and This comprehensively characterizes different degrees and forms of aggregation, such as The value can reflect different degrees of aggregation (mild aggregation, moderate aggregation, severe aggregation). The magnitude of the value can reflect different aggregation patterns (local fiber overlap or large-area aggregation). (2) The fiber length distribution is statistically analyzed using Weibull theory, and a fiber distribution characteristic evaluation model is constructed accordingly, so that the Weibull distribution characteristic parameters obtained at the same time ( and The fiber aggregation morphology on the soil shear surface shows good consistency with the fiber aggregation morphology on the soil shear surface; (3) A fiber distribution characteristic evaluation model based on fiber morphological parameters was established by using the weight allocation method, which improved the rationality and accuracy of the evaluation model. At the same time, the relationship between shear strength and interfacial fiber aggregation index was clarified, which further improved the applicability of the evaluation model. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the fiber in the image acquisition and recognition of the soil shear surface in an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram illustrating the calculation of the fractal dimension of fiber distribution in an image based on the box method in an embodiment of the present invention.

[0019] Figure 3 This is a graph showing the linear regression results based on the least squares method in an embodiment of the present invention.

[0020] Figure 4 This is a schematic diagram of calculating the length of fibers in an image based on the minimum bounding rectangle method in an embodiment of the present invention, where (a) is the extraction of the outer contour of the fiber by findContours, and (b) is the calculation of the minimum bounding rectangle.

[0021] Figure 5 This is a graph showing the results of statistical analysis of fiber length distribution based on Weibull distribution theory in an embodiment of the present invention.

[0022] Figure 6 This is a schematic diagram showing the distribution of statistical parameters of the Weibull distribution in an embodiment of the present invention.

[0023] Figure 7 The figure shows the calculation results and grading evaluation results of the fiber distribution aggregation index on the shear surface obtained by the fiber distribution characteristic evaluation model for evaluating four soil samples in this embodiment of the invention.

[0024] Figure 8 The graph shows the relationship between the shear strength of fiber-reinforced soil and the fiber aggregation index at the shear surface. Detailed Implementation

[0025] The following will provide a clear and complete description of the concept and technical effects of the present invention, so as to fully explain the purpose, solution and effects of the present invention.

[0026] This invention provides a method for quantitatively evaluating the fiber distribution characteristics of soil shear surfaces based on image recognition, which includes the following steps: S1. Obtain an image of the shear plane of the soil, wherein the soil contains fibers; S2. Identify fibers in images based on the double threshold method, and calculate the fractal dimension of fiber distribution based on the box method; S3. Calculate the fiber length in the image based on the minimum bounding rectangle method, and statistically analyze the Weibull distribution characteristic parameters of the fiber length; S4. Based on the fractal dimension and Weibull distribution characteristic parameters, a fiber distribution characteristic evaluation model is established using the weight allocation method. The aggregation index is calculated through the fiber distribution characteristic evaluation model, and the fiber distribution characteristics of the soil shear surface are determined based on the aggregation index.

[0027] The soil sample was red clay containing bamboo fiber. A mixing method was used to prepare the red clay and bamboo fiber, resulting in fiber-reinforced soil. The soil shear surface was obtained by mechanical cutting after a direct shear test (direct shear test), which was performed according to the "Standard for Geotechnical Testing Methods" (GB / T 50123-2019).

[0028] like Figure 1 As shown, a ring light source illuminates the soil shear surface, and an image of the soil shear surface is acquired using a high-definition camera. This image is then processed by a computer. A dual-threshold method is then used to identify fibers on the shear surface, and a green fiber mask is generated using the inRange operation. The specific steps of the dual-threshold method are as follows: The image is preprocessed to grayscale (the size is adjusted to ensure the consistency of the image size when calculating the fiber area ratio, and grayscale processing is performed to obtain an 8-bit grayscale image) to obtain a grayscale image; In a grayscale image, pixels with grayscale values ​​greater than a first preset threshold are set to 255, and pixels with grayscale values ​​less than or equal to the first preset threshold are set to 0. The calculation formula is as follows: ; in, and Two-dimensional coordinates representing image pixels, Corresponding column number (increasing from left to right) Corresponding row numbers (incrementing from top to bottom); This indicates that the first intermediate image generated after binarization according to the first preset threshold is located in the coordinate system. The pixel grayscale value at that location; Represents the grayscale image in coordinates The pixel grayscale value at that location; This represents the first preset threshold (60 in this embodiment); this step can eliminate the interference of low grayscale dark areas in the soil background. Pixels in the grayscale image with grayscale values ​​greater than the second preset threshold are set to 0, and pixels with grayscale values ​​less than or equal to the second preset threshold are set to 255. The calculation formula is as follows: ; in, This indicates that the second intermediate image generated after binarization according to the second preset threshold is located in the coordinate system. The pixel grayscale value at that location; Represents the grayscale image in coordinates The pixel grayscale value at that location; This represents the second preset threshold (120 in this embodiment); this step can remove bright noise in the image, such as reflective spots and bright impurity spots in the soil. Perform a pixel-by-pixel logical AND operation on the first intermediate image and the second intermediate image, retaining only those that simultaneously satisfy the condition "above the first preset threshold". And lower than or equal to the second preset threshold The fiber mask image is obtained by selecting the pixel region (i.e., the bamboo fiber region) based on the given conditions. The calculation formula is as follows: ; in, Represents a fiber mask image. Represents a pixel-by-pixel logical AND operator; if and only if and hour, The result of the operation is 1 (the object being identified is bamboo fiber), otherwise the result is 0 (the object being identified is soil or background). The fibers in the image are determined based on the fiber mask image. The bamboo fiber region in the fiber mask image represents the location of the bamboo fiber in the image of the soil shear surface.

[0029] To remove minute noise from images, morphological processing can be performed using extremely small structural kernels. Specifically, opening operations use a 1×1 elliptic kernel, iterating once; closing operations use a 2×2 elliptic kernel, iterating once. This removes only isolated noise points without damaging the fine fiber structure. An 8-connected component analysis is performed, and a minimum fiber pixel threshold of 1 is set to retain ultrafibers with an area ≥ 1 pixel, thus eliminating regions with minute noise.

[0030] The fiber area percentage on the cut surface is calculated based on the number of identified fibers and the number of pixels. The expression is: ; ; ; in, Represents the total number of pixels in the image. and These represent the length and width of the image, respectively. Represents the total number of pixels in the identified fibers; This indicates an indicator function that takes the value 1 when the condition is met, and 0 otherwise.

[0031] like Figure 2 As shown, the fractal dimension of fiber distribution in an image calculated using the box method includes: converting the fiber mask image into a binary image; constructing squares of different scales with side lengths powers of 2 (e.g., 2, 4, 8, etc.); traversing the image and counting the number of squares containing fibers; taking the natural logarithm of 1 / size and the number of squares; performing a linear regression based on least squares; and the slope of the fitted line is the fractal dimension of the fiber distribution. Figure 3 In this embodiment, 171 fibers were identified; the goodness of fit of the linear regression was 0.99; and the fractal dimension of the fiber distribution on the fitted shear surface was 1.45.

[0032] like Figure 4 As shown, calculating the fiber length in an image based on the minimum bounding rectangle method includes: extracting the fiber outline using findContours and filtering out noise regions with an area less than 10 pixels. Figure 4 (a); then calculate the minimum bounding rectangle for each contour, using the longer side of the rectangle as the effective extension length of the fiber (a); Figure 4(b) Specifically, a circumscribed rectangle with the smallest area is used to completely enclose a single fiber, thereby determining the effective extension length of the fiber. ,in, h i and w i These represent the height and width of the smallest bounding rectangle, respectively.

[0033] Then, the shape parameter and scale parameter of the Weibull distribution are determined according to the following probability density function; ; in, Indicates fiber length. This represents the shape parameter of the Weibull distribution. The scale parameter representing the Weibull distribution (meaning that 63.2% of the fibers are less than [a certain length]). ).like Figure 5 As shown, statistical analysis of fiber length distribution based on Weibull distribution theory is performed to obtain the statistical parameters of the Weibull distribution. and .

[0034] Finally, based on the area proportion Statistical parameters of the Weibull distribution and A fiber distribution characteristic evaluation model was established using a weighted allocation method (with weight coefficients of 0.6, 0.3, and 0.1, respectively), wherein the aggregation index was calculated according to the following formula: ; Among them, AI represents the reunion index. , and Let represent the normalized fractal dimension, the shape parameter of the Weibull distribution, and the scale parameter of the Weibull distribution, respectively. The normalization expression can be: = (D-1.0) / (2-1), =MAX(0.0, MIN(1.0, (1.8- k ) / 1.1)) =MIN(1.0, λ / 30), of which This indicates taking the minimum value. This indicates taking the maximum value.

[0035] Finally, a quantitative evaluation is carried out according to the agglomeration index AI. The specific corresponding relationship between the agglomeration index and the quantitative evaluation is as follows: when AI < 0.45, the conclusion is good dispersion, and the morphological description is uniformly distributed fibers without obvious fiber agglomeration; when 0.45 < AI ≤ 0.65, the conclusion is slight agglomeration, and the morphological description is discrete distribution of fiber clusters with occasional small amounts of agglomerates; when 0.65 < AI ≤ 0.85, the conclusion is moderate agglomeration, and the morphological description is clearly visible fiber bundles with local overlapping regions formed; when AI > 0.85, the conclusion is severe agglomeration, and the morphological description is a large number of fiber bundles with large agglomerates appearing.

[0036] The red clay and bamboo fibers were sampled by the mixing method, and a total of 24 soil samples (fiber-reinforced soil samples) were obtained. They were evaluated respectively by 3 conventional statistical models and the fiber distribution characteristic evaluation model in the embodiment of the present invention. The results are shown in Table 1.

[0037] Table 1 Comparison results of different models Figure 6 It shows a schematic diagram of the distribution of Weibull distribution statistical parameters. It can be seen that there is a certain correlation between whether there is fiber agglomeration on the shear plane and the shape parameters of different agglomeration degrees, that is when > 1.0, there is no agglomeration; when < 1.0, different degrees of agglomeration occur. In addition, there is a certain correlation between the agglomeration morphology and the scale parameter when > 20, the agglomeration morphology shows fiber overlap occurring at multiple places; when < 20, the agglomeration morphology shows a large area of fiber agglomeration concentrated in a certain place.

[0038] Figure 7

[0039] It shows the calculation results of the fiber distribution agglomeration index and the grading evaluation results on the shear plane obtained by evaluating 4 soil samples with the fiber distribution characteristic evaluation model in the embodiment of the present invention. The calculated agglomeration indexes are 0.352, 0.563, 0.741 and 0.921 respectively. According to the specific corresponding relationship between the agglomeration index and the quantitative evaluation, four different morphological fiber distribution characteristics are obtained, namely good dispersion, slight agglomeration, moderate agglomeration and severe agglomeration.

[0039] Figure 8 It shows the measured shear plane strength of the fiber-reinforced soil ( The relationship between fiber distribution characteristics and shear surface fiber aggregation index (AI) shows that as the aggregation index increases, the shear surface strength first increases and then decreases, reflecting the negative impact of fiber aggregation. The goodness of fit of the relationship curve reaches 0.920, indicating that the established fiber distribution characteristic evaluation model has high accuracy and applicability, and can provide basic data and theoretical support for the accurate description of the stress performance of fiber reinforced soil and the establishment of its constitutive model.

[0040] The above description is merely a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. Any embodiment that achieves the technical effects of the present invention by the same or equivalent means should fall within the protection scope of the present invention. Within the protection scope of the present invention, various modifications and variations can be made to the technical solutions and / or implementation methods.

Claims

1. A method for quantitatively evaluating the fiber distribution characteristics of soil shear surfaces based on image recognition, characterized in that, Includes the following steps: Acquire an image of a soil shear plane, wherein the soil contains fibers; The fibers in the image are identified based on the dual threshold method, and the fractal dimension of the fiber distribution is calculated based on the box method. The length of the fibers in the image is calculated based on the minimum bounding rectangle method, and the Weibull distribution characteristic parameters of the fiber length are statistically analyzed. Based on the fractal dimension and Weibull distribution characteristic parameters, a fiber distribution characteristic evaluation model is established using the weight allocation method. The aggregation index is calculated through the fiber distribution characteristic evaluation model, and the fiber distribution characteristics of the soil shear surface are determined based on the aggregation index.

2. The method according to claim 1, characterized in that, The soil is fiber-reinforced soil made from red clay and bamboo fiber using a mixing method.

3. The method according to claim 1, characterized in that, The Weibull distribution characteristic parameters of the statistical fiber length include the shape parameter and scale parameter of the Weibull distribution determined according to the following probability density function; ； in, L Indicates fiber length. k This represents the shape parameter of the Weibull distribution. λ This represents the scale parameter of the Weibull distribution.

4. The method according to claim 1, characterized in that, The formula for calculating the aggregation index is as follows: ； Among them, AI represents the reunion index. , and represents the normalized fractal dimension, the shape parameter of the Weibull distribution, and the scale parameter of the Weibull distribution, respectively.

5. The method according to claim 4, characterized in that, The rules for determining the fiber distribution characteristics of the soil shear surface based on the aggregation index include: If the aggregation index is less than 0.45, the fiber distribution characteristics of the soil shear surface are determined to be well dispersed. If the aggregation index is between 0.45 and 0.65, the fiber distribution characteristics of the soil shear surface are determined to be slight aggregation. If the aggregation index is between 0.65 and 0.85, then the fiber distribution characteristics of the soil shear surface are determined to be moderately aggregated. If the aggregation index is greater than 0.85, the fiber distribution characteristics of the soil shear surface are determined to be severe aggregation.

6. The method according to claim 1, characterized in that, The fractal dimension of fiber distribution calculated using the box method includes: Define box size sequence ,in Indicates the first i The side length of the box, , Indicates the width of the image. Indicates the length of the image. This indicates taking the maximum value. Represents positive integers; For each box size The image is divided into For a given grid size, count the number of fractal boxes that contain at least one fiber pixel. ; Regarding box size and the corresponding number of boxes Take the natural logarithm to obtain the data. The least squares method is used to analyze the data. Perform a linear fit, and the linear fit equation is: ； in, The fractal dimension represents the fiber distribution in the image. This represents the fitted parameters.

7. The method according to claim 1, characterized in that, The method of identifying fibers in the image based on a dual threshold method includes: The image is preprocessed to grayscale to obtain a grayscale image; In the grayscale image, pixels with grayscale values ​​greater than a first preset threshold are set to 255, and pixels with grayscale values ​​less than or equal to the first preset threshold are set to 0, thereby generating a first intermediate image. In the grayscale image, pixels with grayscale values ​​greater than the second preset threshold are set to 0, and pixels with grayscale values ​​less than or equal to the second preset threshold are set to 255, thereby generating a second intermediate image. Perform a pixel-by-pixel logical AND operation on the first intermediate image and the second intermediate image to obtain the fiber mask image; The fibers in the image are determined based on the fiber mask image.

8. The method according to claim 7, characterized in that, The first preset threshold is the minimum gray value of the fiber, and the second preset threshold is the maximum gray value of the fiber.

9. An electronic device, characterized in that, include: At least one processor; and memory that is communicatively connected to at least one processor; The memory stores instructions that, when executed by at least one processor, implement the image recognition-based quantitative assessment method for fiber distribution characteristics of soil shear surfaces as described in any of claims 1-8.

10. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed by a processor, implement the image recognition-based quantitative assessment method for fiber distribution characteristics of soil shear surfaces as described in any of claims 1-8.

Images