A method and system for automatically identifying micro-cracks of concrete and predicting mechanical properties thereof

Abstract

The application discloses a kind of concrete mesoscopic crack automatic identification and mechanical property prediction method and system, belong to concrete identification and prediction technical field, including: obtaining concrete mesoscopic crack image, and crack identification is carried out, crack geometric parameter is extracted based on crack identification result;According to crack geometric parameter, calculate concrete initial damage variable;According to initial damage variable, respectively utilize mesoscopic mechanics homogenization theory and empirical degradation formula, predict the elastic modulus of concrete under current damage state, microcrack friction coefficient, peak damage resistance and uniaxial compressive strength;Further adopt multiscale friction damage constitutive model, generate the stress-strain full curve of concrete under triaxial compression under the damage state.The application uses the above method, realizes the complete prediction from crack image identification to mechanical property, solves the technical problem that initial damage parameter lacks microcosmic foundation.

Description

Technical Field

[0001] This invention relates to the field of concrete identification and prediction technology, and in particular to a method and system for automatic identification of micro-cracks and prediction of mechanical properties in concrete. Background Technology

[0002] Concrete is the most widely used building material in civil engineering, and its mechanical properties and safe life assessment are crucial for the sustainable operation and maintenance of infrastructure. During the pouring and service process, concrete inevitably develops numerous micro-cracks due to factors such as shrinkage, temperature effects, and load application. The presence and development of these cracks directly lead to a decrease in the elastic modulus and strength of the concrete, ultimately causing structural failure. Therefore, quantitatively describing the characteristics of internal cracks in concrete and predicting their impact on mechanical properties is a core task of concrete structural safety assessment.

[0003] In the micromechanics of concrete damage, it is often necessary to define the key parameter of "initial damage variable" to characterize the severity of initial defects in the material. However, existing methods mostly obtain this parameter through empirical setting or back-calculation from macroscopic experiments, lacking quantitative means based on the actual microstructure of the material. On the other hand, while computer vision recognition technology for concrete cracks can automatically extract the geometric information of cracks, it usually ends with image segmentation and parameter statistics, failing to connect with damage mechanics modeling. This disconnect between recognition and modeling means that the rich microscopic information contained in crack images cannot directly serve the performance evaluation of concrete structures.

[0004] Therefore, there is an urgent need in this field for a complete technical solution that can directly convert concrete crack image recognition results into initial damage variables and further predict concrete elastic modulus, strength, and stress-strain curves. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for automatic identification of micro-cracks and prediction of mechanical properties in concrete, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this invention provides a method for automatic identification of micro-cracks and prediction of mechanical properties in concrete, comprising the following steps: S1. Obtain images of fine cracks in concrete, identify cracks in the images, and extract crack geometric parameters based on the crack identification results. S2. Calculate the initial damage variables of the concrete based on the crack geometry parameters; S3. Based on the initial damage variables, the elastic modulus, microcrack friction coefficient, peak damage resistance and uniaxial compressive strength of concrete under the current damage state are predicted using the micromechanical homogenization theory and empirical degradation formula, respectively. S4. Based on the initial damage variables and prediction results, a multi-scale frictional damage constitutive model is used to generate the full stress-strain curve of concrete under triaxial compression under this damage state.

[0007] Preferably, step S1 specifically includes: Acquire images of fine cracks in concrete and convert these images to grayscale to obtain grayscale images. A disk-shaped structuring element is used to perform morphological undercap transformation on grayscale images to extract slender crack features that are darker than the surrounding background, while suppressing large-scale background undulations caused by uneven illumination. Morphological enhancement is performed by dividing the image after morphological cap transformation into several non-overlapping sub-blocks and performing histogram equalization independently within each sub-block to improve the grayscale contrast between the microcracks and the concrete matrix. The enhanced image is subjected to local adaptive threshold segmentation to obtain a preliminary binary image. The preliminary binary image is then cleared of edge pixels and filtered for noise reduction based on the area of ​​connected components to obtain a denoised crack binary image. Skeletonization is performed on the binary image of the crack, the intersection points in the skeleton are detected and dilation is performed to separate the cross crack network into multiple independent branches; Construct a global Euclidean distance field, sample pixel by pixel along the skeleton of each independent branch, extract the physical width of the crack at each position of the branch, and calculate the maximum width and average width. Geometric feature parameters are calculated based on the denoised binary image of the crack and the extracted crack skeleton.

[0008] Preferably, the crack geometry parameters include crack area ratio, number of cracks, total length of crack skeleton, average crack aperture, average crack length, and crack aspect ratio.

[0009] Preferably, step S2 specifically involves: based on the proportion of crack area and the aspect ratio of the crack Using the equivalent medium theory of micromechanics, the initial damage variable of concrete is calculated using the following formula: .

[0010] Preferably, the Mori-Tanaka homogenization method is used for elastic modulus prediction, and its expression is: ; In the formula, Indicates the elastic modulus. and These are the bulk modulus and shear modulus of the concrete matrix, respectively. Represents the damage variable. and To be comparable to Poisson The relevant microcrack influence coefficient is expressed as follows: ; .

[0011] Preferably, the uniaxial compressive strength is calculated based on the exponential decay relationship between the friction coefficient and the peak damage resistance with respect to the damage variable, as shown in the formula: ; ; ; In the formula, and These represent the friction coefficient and peak damage resistance of the microcrack under undamaged conditions. and For decay rate parameters, Peak damage resistance, Indicates the coefficient of friction; according to And calculate the uniaxial compressive strength under the current damage state. : .

[0012] Preferably, the multi-scale frictional damage constitutive model takes the initial damage variables and the predicted microcrack friction coefficient and peak damage resistance as input parameters, and outputs a full stress-strain curve.

[0013] This invention also provides an automatic identification system for micro-cracks in concrete and a system for predicting its mechanical properties, comprising the following modules: Image input module, used to acquire images of micro-cracks in concrete; The crack recognition and geometric parameter extraction module performs grayscale and morphological cap transformation, morphological enhancement, local adaptive threshold segmentation and pixel-level topology repair on the acquired concrete microcrack images in sequence to extract the binary crack image and crack skeleton, and calculate the geometric feature parameters. The initial damage calculation module is used to calculate the initial damage variables of concrete based on geometric feature parameters. The mechanical parameter prediction module is used to predict the elastic modulus, microcrack friction coefficient, peak damage resistance and uniaxial compressive strength of concrete under the current damage state based on the initial damage variables, using the micromechanical homogenization theory and empirical degradation formula respectively. The results output module is used to generate the full stress-strain curve of concrete under triaxial compression conditions in this damage state, based on the initial damage variables and prediction results, using a multi-scale frictional damage constitutive model.

[0014] Therefore, the present invention employs the above-mentioned method and system for automatic identification of micro-cracks and prediction of mechanical properties in concrete, which has the following beneficial effects: (1) It realizes the complete prediction from crack image to mechanical properties, improves the defects of image recognition and mechanical modeling in the existing technology that are independent and cannot be connected, and innovatively combines computer vision technology with micro-damage mechanics theory, opening up the whole path of "image acquisition → crack recognition → damage variable calculation → elastic modulus / strength / stress-strain curve prediction", so that the rich micro information contained in the concrete micro-crack image can directly serve the macro-mechanical performance evaluation. (2) The acquisition of initial damage parameters has a clear microphysical basis. Through the theory of homogenization of micromechanics and empirical degradation formula, the damage variables are calculated directly from the real micro-crack geometric characteristics of the material, which ensures the objectivity and physical interpretability of the damage characterization and improves the shortcomings of existing technologies that rely on empirical presuppositions or macroscopic back deductions, are highly subjective and lack physical basis. (3) It can predict the elastic modulus and uniaxial compressive strength of concrete under different damage states, and generate stress-strain curves, providing key mechanical parameters that are directly available for the assessment of the remaining bearing capacity and prediction of the remaining life of concrete structures under damage states. (4) The accuracy of geometric parameter extraction for complex cross cracks is significantly improved. By adopting the skeleton intersection cutting and distance field sampling method, the cross crack network is actively cut into multiple independent branches for separate analysis. This avoids the deviation in length and width calculation caused by the complexity of skeleton topology in the cross area in the traditional method. The pixel-by-pixel width sampling technology based on distance field can obtain the width distribution of the entire crack. Compared with the average width estimation method of area divided by length, it is more accurate for the characterization of cracks with uneven width and the area near the intersection.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] Figure 1 This is a method diagram of an automatic identification and mechanical property prediction method for micro-cracks in concrete according to an embodiment of the present invention; Figure 2 The images show the recognition results for three typical crack scenarios: (a) is the original simulated image of a single crack; (b) is the original simulated image of a single-point intersecting crack; (c) is the original simulated image of a multi-point intersecting crack; (d) is an enlarged view of the recognition result of a single crack using this method; (e) is an enlarged view of the recognition result of a single-point intersecting crack using this method; and (f) is an enlarged view of the recognition result of a multi-point intersecting crack using this method. Figure 3 The graph shows the degradation curves of elastic modulus under different initial damage states. Figure 4 The graph shows the degradation curves of uniaxial compressive strength under different initial damage states. Figure 5 The friction coefficient degradation curves are shown under different initial damage conditions. Figure 6 The damage resistance degradation curves are shown under different initial damage conditions. Figure 7 This is a stress-strain curve of concrete. Detailed Implementation

[0017] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0019] Example 1 like Figure 1 As shown, this invention provides a method for automatic identification of micro-cracks and prediction of mechanical properties in concrete, comprising the following steps: S1. Obtain microscopic crack images of concrete, perform crack identification on the concrete microscopic crack images, and extract crack geometric parameters based on the crack identification results. Specifically, this includes: Acquire images of microscopic cracks in concrete and convert these images to grayscale to obtain grayscale images. These images can be industrial CT scan slices of concrete specimens, microscopic images of thin sections of concrete taken under an optical microscope or scanning electron microscope, or computer-generated simulated concrete crack images with known crack geometry parameters.

[0020] A disk-shaped structuring element is used to perform morphological undercap transformation on grayscale images to extract slender crack features that are darker than the surrounding background, while suppressing large-scale background undulations caused by uneven illumination. Morphological enhancement is performed by dividing the image after morphological cap transformation into several non-overlapping sub-blocks and performing histogram equalization independently within each sub-block to improve the grayscale contrast between the microcracks and the concrete matrix. The enhanced image is subjected to local adaptive threshold segmentation to obtain a preliminary binary image. The preliminary binary image is then cleared of edge pixels and filtered for noise reduction based on the area of ​​connected components to obtain a denoised crack binary image. Skeletonization is performed on the binary image of the crack, the intersection points in the skeleton are detected and dilation is performed to separate the cross crack network into multiple independent branches; Construct a global Euclidean distance field, sample pixel by pixel along the skeleton of each independent branch, extract the physical width of the crack at each position of the branch, and calculate the maximum width and average width. Based on the denoised binary image of the crack and the extracted crack skeleton, geometric feature parameters are calculated. These parameters include crack area ratio, number of cracks, total length of the crack skeleton, average crack aperture, average crack length, and crack aspect ratio. The number of cracks is calculated by counting the connected components of an 8-connected domain in the skeleton image after removing intersections; the number of connected components is the number of cracks.

[0021] It is important to note that when calculating the total crack length, the length of the central region lost due to the removal of the intersection point needs to be quantitatively compensated. Specifically, an empirical estimation method based on fixed constants is used for quantitative compensation.

[0022] This step enables precise repair of microfractures at crack intersections while maintaining low background noise, ensuring the accuracy of crack count and average length statistics.

[0023] S2. Calculate the initial damage variables of the concrete based on the crack geometry parameters. Specifically: Based on the proportion of crack area and the aspect ratio of the crack Using the equivalent medium theory of micromechanics, the initial damage variable of concrete is calculated using the following formula: ; Among them, the proportion of crack area Corresponding to volume fraction in micromechanics Crack length-to-width ratio Aspect ratio corresponding to microcracks .

[0024] S3. Based on the initial damage variables, the elastic modulus, microcrack friction coefficient, peak damage resistance, and uniaxial compressive strength of concrete under the current damage state are predicted using the micromechanical homogenization theory and empirical degradation formula, respectively.

[0025] The Mori-Tanaka homogenization method is used to predict the elastic modulus, and its expression is as follows: ; In the formula, Indicates the elastic modulus. and These are the bulk modulus and shear modulus of the concrete matrix, respectively. Represents the damage variable. and To be comparable to Poisson The relevant microcrack influence coefficient is expressed as follows: ; .

[0026] Strength prediction is based on friction-damage mechanics theory. This theory posits that the nonlinear mechanical response of quasi-brittle materials under compressive loading is jointly controlled by the frictional slip on the microcrack surface and the damage evolution of the microcrack. The damage evolution of the microcrack follows the following damage criteria: ; The meaning of this criterion is: when the damage driving force of the microcrack... Achieve current damage resistance At that time, the microcracks began to propagate. Damage variables increase, damage driving force It is related to localized stress concentration near the tip of the microcrack.

[0027] Damage resistance function This describes the law governing the material's resistance to microcrack propagation as damage evolves. During loading, the damage resistance exhibits a non-monotonic change characteristic of first increasing and then decreasing: In the initial stage, the microcrack tip needs to overcome a certain energy barrier to initiate cracking, and the damage resistance gradually increases from its initial value; after reaching its peak, the damage resistance gradually decreases as the microcrack penetrates and the material integrity is lost. The form of the damage resistance function is: ; in, Peak damage resistance (i.e., the maximum damage resistance value that the material can reach during loading). The critical damage increment is defined as the damage increment that occurs when the microcrack damage variable reaches its peak resistance. The initial damage deformation is calculated by Module 2.

[0028] Uniaxial compressive strength Based on the theory of friction-damage mechanics, it is predicted that factors such as shrinkage, temperature effects, and loading will cause degradation of the peak damage resistance and friction coefficient on the surface of microcracks inside concrete. A model for peak damage resistance and friction coefficient is established. With damage variables The exponential decay relationship is given by the formula: ; ; In the formula, and These represent the friction coefficient and peak damage resistance of the microcrack under undamaged conditions. and This is the decay rate parameter; Substituting the above attenuation relationship into the microcrack friction-damage strength criterion for concrete, when the confining pressure is 0, we obtain the formula for the degradation of triaxial compressive strength with initial damage: .

[0029] S4. Based on the initial damage variables and prediction results, a multi-scale frictional damage constitutive model is used to generate the full stress-strain curve of concrete under triaxial compression conditions in this damage state. The multi-scale frictional damage constitutive model takes the initial damage variables, microcrack friction coefficient, and peak damage resistance as input parameters and outputs the full stress-strain curve.

[0030] Example 2 This invention also provides an automatic identification system for micro-cracks in concrete and a system for predicting its mechanical properties, comprising the following modules: Image input module, used to acquire images of micro-cracks in concrete; The crack recognition and geometric parameter extraction module performs grayscale and morphological cap transformation, morphological enhancement, local adaptive threshold segmentation and pixel-level topology repair on the acquired concrete microcrack images in sequence to extract the binary crack image and crack skeleton, and calculate the geometric feature parameters. The initial damage calculation module is used to calculate the initial damage variables of concrete based on geometric feature parameters. The mechanical parameter prediction module is used to predict the elastic modulus, microcrack friction coefficient, peak damage resistance and uniaxial compressive strength of concrete under the current damage state based on the initial damage variables, using the micromechanical homogenization theory and empirical degradation formula respectively. The results output module is used to generate the full stress-strain curve of concrete under triaxial compression conditions under the damage state, based on the initial damage variables and prediction results, using a multi-scale frictional damage constitutive model. The multi-scale frictional damage constitutive model takes the initial damage variables, microcrack friction coefficient, and peak damage resistance as input parameters and outputs the full stress-strain curve.

[0031] The specific form of the damage mechanics constitutive model used in the result output generation module of this invention can be selected according to application requirements. Common optional models include damage models based on strain equivalence assumptions, statistical damage models, and friction-damage coupling models based on micromechanics.

[0032] Example 3 Verification of accuracy in simulated crack image recognition: Computer-generated simulated concrete crack images were used as input to verify the accuracy of crack identification and geometric parameter extraction in Module 1 of this method. The simulated images were generated using MATLAB and had a size of 5000×5000 pixels. The following three typical crack scenarios were set: Scenario 1 (Single Crack): A diagonally sloping crack with start and end points at coordinates (2050, 2050) to (2450, 2450), a preset width of 50 pixels, and a simulated image as shown. Figure 2 As shown in (a) in the figure.

[0033] Scenario 2 (Single-point intersecting crack): Two cracks intersect at a single point. Crack 1 has start and end point coordinates from (20500, 2050) to (2450, 2450) and a width of 20 pixels; Crack 2 has start and end point coordinates from (2050, 2450) to (2450, 2050) and a width of 25 pixels. The simulated image is as follows. Figure 2 As shown in (b) of the diagram.

[0034] Scenario 3 (Multi-point Intersecting Cracks): Three cracks form a triangular intersecting network. Crack 1 is a horizontal crack with start and end coordinates from (1900, 2350) to (2500, 2350) and a width of 20 pixels; Crack 2 is a vertical crack with start and end coordinates from (2150, 1900) to (2150, 2600) and a width of 31 pixels; Crack 3 is a 45-degree diagonal through crack with start and end coordinates from (2000, 2000) to (2500, 2500) and a width of 30 pixels. The simulated image is as follows. Figure 2 As shown in (c) in the diagram. Figure 2 (d) in Figure 2 (e) in Figure 2 As shown in (f), this is a magnified view of the recognition results for each scene.

[0035] The simulated images of the three scenarios were input into the crack recognition and geometric parameter extraction module for processing. Binary images and skeletons of the cracks were extracted, and the crack area, number of cracks, average crack width, and average crack length were calculated. The recognition results were compared pixel-by-pixel with the ground truth values, and the errors of each parameter were calculated. The comparison results are summarized in Table 1.

[0036] As shown in Table 1, the area identification errors of this method for single cracks, single-point intersecting cracks, and multi-point intersecting cracks are all controlled within an acceptable range. The crack count identification accuracy is high, and the connectivity of intersection points is well maintained. These results indicate that this method has high accuracy in extracting the geometric parameters of complex crack networks, and can provide a reliable data foundation for the initial damage variable calculation in subsequent module two.

[0037] Table 1. Comparison of initial crack parameter identification and actual conditions under different conditions.

[0038] Example 4 Initial damage variable calculation and mechanical parameter prediction: Taking the recognition result of scenario two (single-point intersection) in Example 3 as an example, the crack area ratio is extracted after processing by the crack recognition and geometric parameter extraction module. and aspect ratio Substituting the numerical value into the initial damage calculation formula, we obtain the initial damage variable under the current damage state. .

[0039] Furthermore, the elastic parameters of the concrete matrix are used: elastic modulus. =150 GPa, Poisson's ratio = 0.27. Based on this, the bulk modulus of the matrix is ​​calculated. and shear modulus and the influence coefficient of microcracks and The initial damage variable Substitute the values ​​into the elastic modulus prediction formula to calculate the elastic modulus E under the current damage state.

[0040] At the same time, the fitting parameters are used , , =0.0138, Substituting -0.541 into the strength prediction formula in Module 3, and using the strength in the undamaged state as a benchmark, the normalized uniaxial compressive strength under the current damaged state is calculated. .

[0041] To demonstrate the degradation patterns of concrete mechanical properties under different damage levels, the initial damage variable was changed. Given the given values, calculate the corresponding elastic modulus and uniaxial compressive strength, and plot the elastic modulus degradation curve and strength deterioration curve, such as... Figure 3 and Figure 4 As shown, the friction coefficient degradation curve and damage resistance degradation curve under different initial damage states are as follows: Figure 5 and Figure 6 As shown in the figure, the results indicate that with the increase of the initial damage variable, both the elastic modulus and the uniaxial compressive strength exhibit a nonlinear decreasing trend, which is consistent with the actual mechanical property degradation law of concrete materials under environmental factors such as freeze-thaw cycles and chemical corrosion.

[0042] Example 5 Generation of concrete stress-strain curves: Based on the different in Example 4 The values ​​are taken and substituted into the multi-scale damage mechanics constitutive model, which automatically generates a set of stress-strain curves covering different initial damage states, such as... Figure 7 As shown, the curve reflects typical damage effects in concrete as initial damage increases, such as a decrease in the slope of the elastic segment, a decrease in peak strength, and an increase in peak strain. It can provide direct mechanical parameter basis for the assessment of the residual bearing capacity and deformation analysis of concrete structures under different damage states.

[0043] Therefore, the present invention employs the above-mentioned method and system for automatic identification of micro-cracks in concrete and prediction of mechanical properties, realizing complete prediction from crack images to mechanical properties, solving the technical problem of lack of micro-foundation for initial damage parameters, and providing an objective and quantifiable technical means for damage assessment and remaining life prediction of concrete structures.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for automatic identification of micro-cracks and prediction of mechanical properties in concrete, characterized in that, Includes the following steps: S1. Obtain images of fine cracks in concrete, identify cracks in the images, and extract crack geometric parameters based on the crack identification results. S2. Calculate the initial damage variables of the concrete based on the crack geometry parameters; S3. Based on the initial damage variables, the elastic modulus, microcrack friction coefficient, peak damage resistance and uniaxial compressive strength of concrete under the current damage state are predicted using the micromechanical homogenization theory and empirical degradation formula, respectively. S4. Based on the initial damage variables and prediction results, a multi-scale frictional damage constitutive model is used to generate the full stress-strain curve of concrete under triaxial compression under this damage state.

2. The method for automatic identification of micro-cracks and prediction of mechanical properties in concrete according to claim 1, characterized in that, Step S1 specifically includes: Acquire images of fine cracks in concrete and convert these images to grayscale to obtain grayscale images. A disk-shaped structuring element is used to perform morphological undercap transformation on grayscale images to extract slender crack features that are darker than the surrounding background, while suppressing large-scale background undulations caused by uneven illumination. Morphological enhancement is performed by dividing the image after morphological cap transformation into several non-overlapping sub-blocks and performing histogram equalization independently within each sub-block to improve the grayscale contrast between the microcracks and the concrete matrix. The enhanced image is subjected to local adaptive threshold segmentation to obtain a preliminary binary image. The preliminary binary image is then cleared of edge pixels and filtered for noise reduction based on the area of ​​connected components to obtain a denoised crack binary image. Skeletonization is performed on the binary image of the crack, the intersection points in the skeleton are detected and dilation is performed to separate the cross crack network into multiple independent branches; Construct a global Euclidean distance field, sample pixel by pixel along the skeleton of each independent branch, extract the physical width of the crack at each position of the branch, and calculate the maximum width and average width. Geometric feature parameters are calculated based on the denoised binary image of the crack and the extracted crack skeleton.

3. The method for automatic identification of micro-cracks and prediction of mechanical properties in concrete according to claim 2, characterized in that: Crack geometry parameters include crack area percentage, number of cracks, total length of crack skeleton, average crack aperture, average crack length, and crack aspect ratio.

4. The method for automatic identification of micro-cracks and prediction of mechanical properties in concrete according to claim 3, characterized in that, Step S2 specifically involves: based on the proportion of crack area and the aspect ratio of the crack Using the equivalent medium theory of micromechanics, the initial damage variable of concrete is calculated using the following formula: 。 5. The method for automatic identification of micro-cracks and prediction of mechanical properties in concrete according to claim 1, characterized in that, The Mori-Tanaka homogenization method is used to predict the elastic modulus, and its expression is as follows: ； In the formula, Indicates the elastic modulus. and These are the bulk modulus and shear modulus of the concrete matrix, respectively. Represents the damage variable. and To be comparable to Poisson The relevant microcrack influence coefficient is expressed as follows: ； 。 6. The method for automatic identification of micro-cracks and prediction of mechanical properties in concrete according to claim 1, characterized in that: Uniaxial compressive strength is calculated based on the friction coefficient and the exponential decay relationship of peak damage resistance with respect to damage variables, as shown in the formula: ； ； ； In the formula, and These represent the friction coefficient and peak damage resistance of the microcrack under undamaged conditions. and For decay rate parameters, Peak damage resistance, Indicates the coefficient of friction; according to And calculate the uniaxial compressive strength under the current damage state. : 。 7. The method for automatic identification of micro-cracks and prediction of mechanical properties in concrete according to claim 6, characterized in that: The multi-scale frictional damage constitutive model takes the initial damage variables, the predicted microcrack friction coefficient, and the peak damage resistance as input parameters and outputs a full stress-strain curve.

8. An automatic identification system for micro-cracks in concrete and a system for predicting its mechanical properties, comprising the method for automatic identification of micro-cracks in concrete and prediction of its mechanical properties as described in any one of claims 1-7, characterized in that, Includes the following modules: Image input module, used to acquire images of micro-cracks in concrete; The crack recognition and geometric parameter extraction module performs grayscale and morphological cap transformation, morphological enhancement, local adaptive threshold segmentation and pixel-level topology repair on the acquired concrete microcrack images in sequence to extract the binary crack image and crack skeleton, and calculate the geometric feature parameters. The initial damage calculation module is used to calculate the initial damage variables of concrete based on geometric feature parameters. The mechanical parameter prediction module is used to predict the elastic modulus, microcrack friction coefficient, peak damage resistance and uniaxial compressive strength of concrete under the current damage state based on the initial damage variables, using the micromechanical homogenization theory and empirical degradation formula respectively. The results output module is used to generate the full stress-strain curve of concrete under triaxial compression conditions in this damage state, based on the initial damage variables and prediction results, using a multi-scale frictional damage constitutive model.

Images