Method for evaluating appearance quality of highway engineering precast beam concrete

Abstract

The application provides a highway engineering prefabricated beam concrete appearance quality evaluation method, which collects the image of the prefabricated beam concrete through machine vision, carries out pretreatment through MATLAB software, then carries out neural network training, and then identifies the corresponding holes, constructs a scoring model according to the area ratio of the holes to the image area, so that accurate scoring results can be obtained, the appearance image information of the prefabricated beam is obtained through machine vision, then appearance scoring is carried out based on the image information, so that the appearance quality evaluation of the prefabricated beam is good, and the accuracy of the final evaluation is ensured, the influence of subjective factors in the traditional technology is effectively avoided, and the final quality of the highway engineering is ensured.

Description

Technical Field

[0001] This invention relates to a method for evaluating the quality of concrete, and more particularly to a method for evaluating the appearance quality of precast concrete beams in highway engineering. Background Technology

[0002] Precast beams are widely used in existing highway projects, and the quality evaluation of the concrete in these precast beams significantly affects the overall quality of the highway project.

[0003] In existing technologies, the evaluation of the appearance of precast beams is only qualitative, that is, the quality of precast beams is qualitatively evaluated by experienced personnel. This results in the evaluation process being greatly affected by subjective factors such as the experience of the personnel, which seriously affects the accurate evaluation of the appearance quality of precast beams and creates safety hazards for highway engineering.

[0004] Therefore, in order to solve the above-mentioned technical problems, it is urgent to propose a new technical approach. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method for evaluating the appearance quality of precast concrete beams in highway engineering. This method acquires appearance image information of precast beams through machine vision, and then scores the appearance based on the image information, thereby effectively evaluating the appearance quality of precast beams and ensuring the accuracy of the final assessment. This effectively avoids the influence of subjective factors in traditional technologies and ensures the final quality of highway engineering.

[0006] This invention provides a method for evaluating the appearance quality of precast concrete beams in highway engineering, comprising the following steps:

[0007] S1. Acquire image information of the target precast beam concrete;

[0008] S2. Input the image information of the concrete into MATLAB software for preprocessing;

[0009] S3. Use the round function to find the largest integers whose length and width are divisible by 10 after preprocessing the image;

[0010] S4. Use the imresize function to adjust the image length and width to the calculated maximum integers m and n, where m represents the maximum integer length divisible by 10, and n is the maximum integer width divisible by 10.

[0011] S5. Divide the image processed in step S3 into 100 equal parts, with each part having an image size of [size missing]. Then, the imresize function is used to resize each of the evenly divided images to 360*360.

[0012] S6. Use the Image Labeler tool to perform hole annotation on the 100 images processed in step S5, and use the annotated images as the training set;

[0013] S7. Input the images from the training set into the neural network;

[0014] The neural network is trained by calling the training set using the trainFasterRCNNObjectDetector function.

[0015] S8. Use the trained neural network to identify holes in each image after processing in step S3, and use the insertShape function to mark the location of the identified holes and determine the area ratio of the holes in the corresponding image.

[0016] S9. Construct an appearance quality scoring model W:

[0017] Where k is the percentage of the hole area.

[0018] Furthermore, in step S8, the area ratio of the holes is determined using the following method:

[0019] The insertShape function uses a rectangle to mark the location of the hole;

[0020] Determine the length 'a' and width 'b' of the rectangle;

[0021] Construct a model of the proportion of hole area:

[0022] Where β is the hole area ratio coefficient and s is the area of ​​the image.

[0023] Furthermore, MATLAB software performs image preprocessing including:

[0024] The background image is extracted from the image using a Gaussian low-pass filter;

[0025] Subtract the brightness of the background image from that of the original image;

[0026] The original image after brightness subtraction is stretched in grayscale.

[0027] The beneficial effects of this invention are as follows: By using machine vision to acquire the appearance image information of precast beams, and then performing appearance scoring based on the image information, the appearance quality of precast beams can be effectively evaluated, thereby ensuring the accuracy of the final assessment, effectively avoiding the influence of subjective factors in traditional technologies, and ensuring the final quality of highway engineering. Attached Figure Description

[0028] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0029] Figure 1 This is a flowchart of the present invention.

[0030] Figure 2 This is a comparison image of the original image preprocessing in this invention. Detailed Implementation

[0031] The present invention will be further described in detail below:

[0032] This invention provides a method for evaluating the appearance quality of precast concrete beams in highway engineering, comprising the following steps:

[0033] S1. Acquire image information of the target precast beam concrete;

[0034] S2. Input the image information of the concrete into MATLAB software for preprocessing;

[0035] S3. Use the round function to find the largest integers whose length and width are divisible by 10 after preprocessing the image;

[0036] S4. Use the imresize function to adjust the image length and width to the calculated maximum integers m and n, where m represents the maximum integer length divisible by 10, and n is the maximum integer width divisible by 10.

[0037] S5. Divide the image processed in step S3 into 100 equal parts, with each part having an image size of [size missing]. Then, the imresize function is used to resize each of the evenly divided images to 360*360.

[0038] S6. Use the Image Labeler tool to perform hole annotation on the 100 images processed in step S5, and use the annotated images as the training set;

[0039] S7. Input the images from the training set into the neural network;

[0040] The neural network is trained by calling the training set using the trainFasterRCNNObjectDetector function.

[0041] S8. Use the trained neural network to identify holes in each image after processing in step S3, and use the insertShape function to mark the location of the identified holes and determine the area ratio of the holes in the corresponding image.

[0042] S9. Construct an appearance quality scoring model W:

[0043] Where k represents the proportion of the hole area. Using the above method, the appearance image information of the precast beam is acquired through machine vision, and then appearance scoring is performed based on the image information, thus providing a good evaluation of the appearance quality of the precast beam. This ensures the accuracy of the final assessment, effectively avoids the influence of subjective factors in traditional techniques, and guarantees the final quality of the highway project. All of the above processes are executed in MATLAB software.

[0044] In this embodiment, in step S8, the area ratio of the holes is determined by the following method:

[0045] The insertShape function uses a rectangle to mark the location of the hole;

[0046] Determine the length 'a' and width 'b' of the rectangle;

[0047] Construct a model of the proportion of hole area:

[0048] Where β is the hole area ratio coefficient, which is determined according to the actual situation; p is the number of holes in each image; generally, the value is 0.8; s is the area of ​​the image. Then, after each original image is processed by steps S2-S5, each image has 100 copies, so s = 360 * 360 * 100.

[0049] In this embodiment, the image preprocessing using MATLAB software includes: preprocessing using the MASK homogenization method, specifically including:

[0050] The background image is extracted from the image using a Gaussian low-pass filter;

[0051] The brightness of the background image is subtracted from that of the original image, thus making the brightness contrast of the image more uniform.

[0052] The original image, after brightness subtraction, undergoes grayscale stretching to increase the contrast between adjacent details. This method improves the accuracy of the final hole identification result.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. 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 be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for evaluating the appearance quality of precast concrete beams in highway engineering, characterized in that: Includes the following steps: S1. Acquire image information of the target precast beam concrete; S2. Input the image information of the concrete into MATLAB software for preprocessing; S3. Use the round function to find the largest integers whose length and width are divisible by 10 after preprocessing the image; S4. Use the imresize function to adjust the image length and width to the calculated maximum integers m and n, where m represents the maximum integer length divisible by 10, and n is the maximum integer width divisible by 10. S5. Divide the image processed in step S3 into 100 equal parts, with each part having an image size of [size missing]. Then, the imresize function is used to resize each of the evenly divided images to 360*360. S6. Use the Image Labeler tool to perform hole annotation on the 100 images processed in step S5, and use the annotated images as the training set; S7. Input the images from the training set into the neural network; The neural network is trained by calling the training set using the trainFasterRCNNObjectDetector function. S8. Use the trained neural network to identify holes in each image after processing in step S3, and use the insertShape function to mark the location of the identified holes and determine the area ratio of the holes in the corresponding image. S9. Construct an appearance quality scoring model W: Where k is the percentage of the hole area.

2. The method for evaluating the appearance quality of precast concrete beams in highway engineering according to claim 1, characterized in that: In step S8, the area ratio of the holes is determined using the following method: The insertShape function uses a rectangle to mark the location of the hole; Determine the length 'a' and width 'b' of the rectangle; Construct a model of the proportion of hole area: Where β is the hole area ratio coefficient and s is the area of ​​the image.

3. The method for evaluating the appearance quality of precast concrete beams in highway engineering according to claim 1, characterized in that: MATLAB software performs image preprocessing including: The background image is extracted from the image using a Gaussian low-pass filter; Subtract the brightness of the background image from that of the original image; The original image after brightness subtraction is stretched in grayscale.

Images