Fine crack assessment method for cement-based materials based on digital image processing

Abstract

The embodiment of the present application provides a method for evaluating fine cracks in cement-based materials based on digital image processing. The technical key points are: stretching a target specimen of a high-toughness cement-based composite material ECC, collecting a crack image once at a fixed interval through a camera, and obtaining a background-corrected crack image by performing signal averaging conversion, median filter filtering and other operations on the crack image; obtaining qualified binary images that accurately contain information on fine cracks and wide cracks in sequence through an edge recognition algorithm or a dual threshold algorithm; denoising and enhancing each of the qualified binary images in reverse and forward order according to the time sequence axis of shooting, obtaining the pixel area and boundary coordinates of each crack region through each of the denoised images, identifying the crack width as the projection length of the crack in the loading direction, and obtaining the average width of the crack by dividing the pixel area of ​​the crack region by the crack length, accumulating the crack length and dividing it by the entire width of the specimen to obtain the number of cracks.

Description

Technical Field

[0001] The present application relates to the field of concrete structure crack detection, and in particular to a cement-based material fine crack assessment method based on digital image processing. Background Art

[0002] Cement-based materials are a large-scale and widely used engineering material. The performance of cement-based materials in various forms of structures is seriously degraded due to cracking under the long-term coupling of environment and load. Accurate detection and identification of cracks are of great practical significance. At present, ultra-high toughness fiber-reinforced cement-based composites (ECC) can show fine multi-crack cracking under tensile and shear loads. The crack width can be controlled below 60μm, and the crack spacing can reach the millimeter level. High-precision and easily available digital images are required.

[0003] In the prior art, the ECC crack detection method based on digital image recognition mainly adopts the digital image correlation method (DIC) and the digital image processing method (DIP). The industrial camera and supporting computing software required by the DIC method are relatively expensive and cannot be widely used. At the same time, since ECC cracks are relatively fine, the size of the crack sub-areas divided during DIC processing is small, resulting in low image processing efficiency. The traditional DIP method uses a single domain value algorithm to identify ECC cracks, and the crack identification effect is poor, and it is impossible to take into account both low noise and high-quality crack information. Summary of the invention

[0004] The embodiment of the present application provides a method for evaluating fine cracks in cement-based materials based on digital image processing, so as to at least solve the problem in the related art that crack recognition effect is poor and it is impossible to take into account both low noise and high-quality crack information.

[0005] In one embodiment of the present application, a method for evaluating fine cracks in cement-based materials based on digital image processing is proposed, wherein the method comprises:

[0006] The target specimen of high-toughness cement-based composite material ECC was stretched, and the camera was used to collect crack images at fixed intervals. When the main crack of the target specimen was formed, the camera was used to continuously collect crack images, where each photo was separated by a distance of one pixel.

[0007] The crack image is converted into a grayscale image by signal averaging, the crack structure on the grayscale image is filtered by a median filter to obtain a smooth image, and the background-corrected crack image is obtained by subtracting the smooth image from the grayscale image;

[0008] A qualified binary image containing accurate information of fine cracks and wide cracks is obtained in sequence through an edge recognition algorithm or a double threshold algorithm;

[0009] According to the time sequence axis of shooting, each of the qualified binary images is subjected to denoising and enhancement processing in reverse order to obtain a pre-processed image, and the pre-processed image and the qualified binary image are subjected to pixel matrix multiplication denoising processing to obtain a denoised image;

[0010] The pixel area and boundary coordinates of each crack region are obtained through each of the denoised images. The crack width is identified as the projected length of the crack in the loading direction. The average width of the crack is obtained by dividing the pixel area of ​​the cracked region by the crack length. The crack length is accumulated and divided by the entire width of the specimen to obtain the number of cracks.

[0011] In one embodiment, the denoising and enhancement processing is performed on each of the qualified binary images in reverse order forward according to the time sequence axis of shooting, including: using image processing software to manually remove noise points around cracks in each binary image of the last frame on the time sequence axis.

[0012] In one embodiment, the crack image includes a plurality of continuously shot images of the final state of stretching, and the plurality of continuously shot images are synthesized into a photo with a higher pixel density using image synthesis software or drizzle algorithm.

[0013] In one embodiment, the background-corrected crack image is converted into a low binary image using a low threshold;

[0014] Using a high threshold to convert the background-corrected crack image into a high binary image;

[0015] Subtracting the low binary image from the high binary image to obtain a high-retention image that effectively retains fine crack information and wide crack contour information;

[0016] Performing edge enhancement on the high binary image based on a Hessian matrix and a Gaussian filter to obtain a high binary image of a wide crack with a widened contour;

[0017] Performing a pixel matrix multiplication operation on the high-retention image and the high-binary image of the wide crack with a widened contour, removing fine crack information contained in the high-retention image, and obtaining a high-binary image that only retains the contour information of the wide crack;

[0018] The high binary image retaining only the wide crack contour information is subtracted from the high binary image to cut off the redundant wide crack contours of the high binary image, thereby obtaining a qualified binary image containing accurate wide crack information and fine crack information.

[0019] In one embodiment, Gaussian-Laplace operator is used to perform Gaussian filtering on the background-corrected crack image to remove significant noise effects;

[0020] The Gaussian-Laplace operator is used to perform secondary processing on the image to improve the ability to extract the image edge;

[0021] The obtained edges are enhanced, the crack width is calculated by the number of pixels, and the calculated results are compared with the microscope observation results to adjust the image algorithm processing parameters.

[0022] Through the embodiments of the present application, compared with the problem of crack information distortion when the single domain value method binarizes the crack grayscale image, the double threshold method can well retain the fine crack information and wide crack information; the crack photos taken by ordinary cameras, due to the limitation of pixel density, will cause small cracks to be too small to be identified because of their width. Using image synthesis software, multiple ordinary precision photos can be synthesized into a high-precision photo, and small cracks can also be well identified; the crack width and number change law of the whole ECC stretching process can be obtained, and the tensile strain level corresponding to the given crack width limit can be effectively given, which is of great significance for evaluating the permeability and durability of materials; the edge recognition algorithm or double threshold algorithm used has lower requirements on computer equipment, and can effectively improve the calculation efficiency without reducing the accuracy compared to the DIC algorithm. It solves the problem of high cost and low efficiency of the existing DIC method for evaluating the fine cracking of ECC materials, and the problem of crack information distortion of the traditional DIP evaluation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0024] Figure 1 is a flow chart according to an embodiment of the present application;

[0025] Figure 2 is a high pixel density crack map according to an embodiment of the present application;

[0026] Figure 3 is a grayscale crack map according to an embodiment of the present application;

[0027] Figure 4 is a 0.35 threshold crack binary image according to an embodiment of the present application;

[0028] Figure 5 is a crack binary image with a threshold value of 0.53 according to an embodiment of the present application;

[0029] Figure 6 is a flow chart of a dual threshold image processing algorithm according to an embodiment of the present application;

[0030] Figure 7is a crack distribution diagram after processing according to an embodiment of the present application;

[0031] Figure 8 This is the identification of crack width according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0033] It should be noted that the terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0034] like Figure 1 As shown, a method for evaluating fine cracks in cement-based materials based on digital image processing is proposed, and the method includes:

[0035] S202, stretching a target specimen of high-toughness cement-based composite material ECC, collecting crack images once at fixed intervals through a camera, and continuously collecting crack images through a camera when a main crack is formed near the target specimen, wherein each photo is separated by a distance of one pixel;

[0036] S204, converting the crack image into a grayscale image by signal averaging, filtering the crack structure on the grayscale image with a median filter to obtain a smoothed image, and subtracting the smoothed image from the grayscale image to obtain a background-corrected crack image;

[0037] S206, sequentially obtaining qualified binary images that accurately contain information on fine cracks and wide cracks through an edge recognition algorithm or a dual threshold algorithm;

[0038] S208, performing denoising and enhancement processing on each of the qualified binary images in reverse order and forward order according to the shooting time sequence axis to obtain a pre-processed image, and performing pixel matrix multiplication denoising processing on the pre-processed image and the qualified binary image to obtain a denoised image;

[0039] S210, the pixel area and boundary coordinates of each crack region are obtained through each of the denoised images, the crack width is identified as the projection length of the crack in the loading direction, the average width of the crack is obtained by dividing the pixel area of ​​the crack region by the crack length, the crack length is accumulated and divided by the entire width of the specimen to obtain the number of cracks.

[0040] It should be noted that this application is aimed at the poor crack recognition effect, which cannot take into account both low noise and high-quality crack information. Ordinary cameras can be used to analyze the number and width of ECC cracks under any strain, and edge recognition algorithms or dual threshold algorithms can be used to improve the crack recognition effect. At the same time, denoising algorithms and crack separation algorithms are used to obtain the development process of any crack under any load level, which can be used to evaluate the cracking performance of ECC materials. The mature characterization and identification methods for ECC cracks focus on comparing microscope values ​​with algorithm results.

[0041] In one embodiment, the denoising and enhancing processing is performed on each of the qualified binary images in reverse order according to the time sequence of shooting, and further includes:

[0042] S1, image processing software was used to manually remove the noise around the cracks of each binary image of the last frame on the time sequence axis.

[0043] It should be noted that manual noise reduction can be performed using the commercial software Photoshop.

[0044] In one embodiment, the crack image includes a plurality of continuously shot images of the final state of stretching, and the plurality of continuously shot images are synthesized into a photo with a higher pixel density using image synthesis software or drizzle algorithm.

[0045] In one embodiment, the dual threshold algorithm comprises the following steps:

[0046] S1, using a low threshold to convert the background corrected crack image into a low binary image; at this time, the image processed by the low threshold can effectively retain the fine crack information, but the wide crack information will be distorted due to the enlargement of the edge contour.

[0047] S2, converting the background-corrected crack image into a high binary image using a high threshold; the grayscale image processed by the high threshold can accurately obtain wide crack information, but ignores fine crack information;

[0048] S3, subtracting the low binary image from the high binary image to obtain a high-retention image that effectively retains fine crack information and wide crack contour information;

[0049] S4, performing edge enhancement on the high binary image based on a Hessian matrix and a Gaussian filter to obtain a high binary image of a wide crack with a widened contour;

[0050] S5, performing a pixel matrix multiplication operation on the high-retention image and the high-binary image of the wide crack with a widened contour, removing the fine crack information contained in the high-retention image, and obtaining a high-binary image that only retains the contour information of the wide crack;

[0051] S6, subtracting the high binary image retaining only the wide crack contour information from the high binary image to cut off the redundant wide crack contours of the high binary image, and obtaining a qualified binary image containing accurate wide crack information and fine crack information.

[0052] In one embodiment, the edge recognition dual threshold algorithm comprises the following steps:

[0053] S1, using Gaussian-Laplace operator to perform Gaussian filtering on the background-corrected crack image to remove more significant noise effects;

[0054] S2, using Gaussian-Laplace operator to perform secondary processing on the image to improve the ability to extract image edges;

[0055] S3, strengthening the obtained edge, calculating the crack width by the number of pixels, comparing the calculated result with the microscope observation result, and adjusting the image algorithm processing parameters.

[0056] like Figures 2 to 8 As shown, the following specific example is used to explain:

[0057] 1. As Figure 2 As shown, the present invention uses a Sony 90 macro lens to shoot a test piece with a surface area of ​​30 mm × 45 mm, and synthesizes 16 photos with a pixel number of 1616 × 1080 pixels into a photo with a pixel number of 12672 × 19008. The pixel number of the selected area is 12672 × 8424, and the RGB image is converted into a grayscale image, as shown in Figure 3 shown.

[0058] 2. The image is processed with a threshold of 0.35 to effectively retain the wide crack information, such as Figure 4 shown.

[0059] 3. The grayscale image is processed with a threshold of 0.53. At this time, wide cracks will be widened and fine cracks will appear, such as Figure 5 shown.

[0060] 4. By subtracting the two, an image with both fine crack and coarse crack contours can be obtained.

[0061] 5. The coarse crack image is enhanced using a Gaussian filter. The Gaussian filter window height and width are selected as 3, and the Sigma value of the Gaussian convolution is 1.

[0062] 6. Multiplying the images obtained in 4 and 5 can obtain the coarse crack contour information and ignore the fine crack information.

[0063] 7. Subtracting the image in 3 from the image in 6 can obtain a more accurate image of the coexistence of coarse cracks and fine cracks. The flowchart of the double threshold method is as follows: Figure 5 shown.

[0064] 8. The image now contains some noise. Manual noise reduction is performed using commercial software Photoshop. The image processing results are as follows: Figure 6 .

[0065] 9. The crack width can be identified based on the number of pixels, e.g. Figure 7 As shown, the crack occupies 28 pixels and its width is 66.3 μm.

[0066] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for evaluating fine cracks in cement-based materials based on digital image processing, characterized in that: The method comprises: The target specimen of high-toughness cement-based composite material ECC was stretched, and the camera was used to collect crack images at fixed intervals. When the main crack of the target specimen was formed, the camera was used to continuously collect crack images, where each photo was separated by a distance of one pixel. The crack image is converted into a grayscale image by signal averaging, the crack structure on the grayscale image is filtered by a median filter to obtain a smooth image, and the background-corrected crack image is obtained by subtracting the smooth image from the grayscale image; A qualified binary image containing accurate information of fine cracks and wide cracks is obtained in sequence through an edge recognition algorithm or a double threshold algorithm; According to the time sequence axis of shooting, each of the qualified binary images is subjected to denoising and enhancement processing in reverse order to obtain a pre-processed image, and the pre-processed image and the qualified binary image are subjected to pixel matrix multiplication denoising processing to obtain a denoised image; The pixel area and boundary coordinates of each crack region are obtained through each of the denoised images. The crack width is identified as the projected length of the crack in the loading direction. The average width of the crack is obtained by dividing the pixel area of ​​the cracked region by the crack length. The crack length is accumulated and divided by the entire width of the specimen to obtain the number of cracks.

2. The method according to claim 1, characterized in that The denoising and enhancing processes are performed on each of the qualified binary images in reverse and forward order according to the time sequence axis of shooting, including: using image processing software to manually remove noise points around cracks in each grid of the last binary image on the time sequence axis.

3. The method according to claim 1, characterized in that The crack image includes a plurality of continuously shot images of the final stretching state, and the plurality of continuously shot images are synthesized into a photo with a higher pixel density by using image synthesis software or drizzle algorithm.

4. The method according to claim 1, characterized in that: The dual threshold algorithm comprises the following steps: Converting the background-corrected crack image into a low binary image using a low threshold; The background-corrected crack image is converted into a high binary image by using a high threshold; the low binary image is subtracted from the high binary image to obtain a high-retention image that effectively retains fine crack information and wide crack contour information; Performing edge enhancement on the high binary image based on a Hessian matrix and a Gaussian filter to obtain a high binary image of a wide crack with a widened contour; Performing a pixel matrix multiplication operation on the high-retention image and the high-binary image of the wide crack with a widened contour, removing fine crack information contained in the high-retention image, and obtaining a high-binary image that only retains the contour information of the wide crack; The high binary image retaining only the wide crack contour information is subtracted from the high binary image to cut off the redundant wide crack contours of the high binary image, thereby obtaining a qualified binary image containing accurate wide crack information and fine crack information.

5. The method according to claim 1, characterized in that The edge recognition algorithm comprises the following steps: The Gaussian-Laplace operator is used to perform Gaussian filtering on the background-corrected crack image to remove significant noise effects; The Gaussian-Laplace operator is used to perform secondary processing on the image to improve the ability to extract the image edge; The obtained edges are enhanced, the crack width is calculated by the number of pixels, and the calculated results are compared with the microscope observation results to adjust the image algorithm processing parameters.

Images