Bituminous pavement repairing quality intelligent evaluation method and system based on machine vision

Abstract

The invention relates to a machine vision-based intelligent evaluation method and system for the repairing quality of an asphalt pavement, and the method comprises the steps: collecting a reference image before the repairing of the asphalt pavement and a comparison image after the repairing, and carrying out the defect feature extraction of the reference image and the comparison image, the defect area and the defect strength of the asphalt pavement before repairing and the defect area and the defect strength of the asphalt pavement after repairing are extracted respectively, and the repairing effect can be evaluated by comparing defect characteristics. The evaluation dimensions are respectively defect area and defect intensity. Brightness unification is carried out based on the standard gray scale card, so that the light problem can be overcome. Besides, feature extraction is performed on the asphalt pavement before and after repairing, and the defect features further comprise defect strength, so that the repairing effect of the asphalt pavement can be evaluated more accurately.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to a machine vision-based intelligent assessment method and system for asphalt pavement repair quality. Background Technology

[0002] Currently, the quality of asphalt pavement repairs is mainly evaluated based on manual experience. However, this reliance on manual methods has significant drawbacks, primarily low efficiency and inconsistent standards. Therefore, machine vision is increasingly being incorporated into existing technologies to assess the quality of asphalt pavement repairs.

[0003] Asphalt pavements are typically used outdoors, and their lighting conditions are drastically affected by weather (sunny, cloudy, rainy, snowy), time of day (morning / evening, midday), and shadows (trees, buildings). Uneven lighting can lead to decreased image contrast and color distortion, affecting the identification of defects such as cracks and potholes. Furthermore, existing machine vision-based assessment methods generally only evaluate the quality of the repaired pavement, but because the repaired pavement still has a color difference from the original pavement, misassessment is prone to occur. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a machine vision-based intelligent assessment method and system for asphalt pavement repair quality, so as to solve the above-mentioned technical problems.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] The intelligent assessment method for asphalt pavement repair quality based on machine vision of the present invention includes the following steps:

[0007] Acquire a reference image before road surface repair and a comparison image after road surface repair, wherein both the reference image and the comparison image contain a standard grayscale card area;

[0008] The reference image and the comparison image are preprocessed based on the standard grayscale card area to obtain a preprocessed reference image and a preprocessed comparison image. The preprocessing includes grayscale conversion, filtering, and brightness unification.

[0009] The reference defect region and the reference defect feature of the reference defect region are extracted from the preprocessed reference image, and the comparison defect region and the comparison defect feature of the comparison defect region are extracted from the preprocessed comparison image. The reference defect region and the comparison defect region both include crack regions and pit regions, and the reference defect feature and the comparison defect feature both include area features and intensity features.

[0010] The baseline defect features of the baseline defect region are compared with the comparative defect features of the comparative defect region to obtain a comparison result; and evaluation information is generated based on the comparison result.

[0011] In one embodiment of this application, preprocessing is performed on the reference image and the comparison image based on the standard grayscale card region to obtain a preprocessed reference image and a preprocessed comparison image, including:

[0012] The image to be processed is converted to grayscale and high-pass filtered to obtain a grayscale image, wherein the image to be processed is a reference image or a comparison image;

[0013] Extract the grayscale values ​​of multiple pixels in the standard grayscale card area, and calculate the average of the grayscale values ​​of the multiple pixels to obtain the grayscale reference value gray. ref ;

[0014] Obtain the true grayscale value of a standard grayscale card. real And calculate the grayscale reference value gray. ref Compared with the actual gray value gray real The ratio k, k = gray real / gray ref Based on the ratio, the grayscale image is subjected to brightness unification processing to obtain a preprocessed image, wherein the preprocessed image is a preprocessing reference image or a preprocessing comparison image, and the mathematical expression of the preprocessed image is:

[0015] preprocessing(i,j)=k×gray(i,j)

[0016] In the formula, preprocessing(i,j) represents the gray value of pixel (i,j) in the preprocessed image, and gray(i,j) represents the gray value of pixel (i,j) in the grayscale image.

[0017] In one embodiment of this application, extracting a reference defect region and reference defect features of the reference defect region from the preprocessed reference image, and extracting a comparison defect region and comparison defect features of the comparison defect region from the preprocessed comparison image, includes:

[0018] The preprocessed image is then meshed to obtain a meshed image. grid (x,y), where (x,y) represents grid coordinates, and the preprocessing image is a preprocessing reference image or a preprocessing comparison image;

[0019] Calculate the gridded imagegrid Road surface texture features for each grid texture The road surface texture feature texture Includes grayscale mean gray(x,y) and multi-directional average GLCM entropy E avg (x,y);

[0020] Grids that meet the target conditions are used as the target grid. target The target conditions include a mean grayscale value greater than a preset grayscale threshold and a multi-directional average GLCM entropy E. avg (x,y) is less than or equal to a preset entropy threshold;

[0021] Calculate the target grid target The Hessian features of the inner pixels are used to locate the crack region based on the Hessian features;

[0022] Extract the pitted regions from the preprocessed image;

[0023] Extract the defect features of the crack region and the defect features of the pit region.

[0024] In one embodiment of this application, the gridded image is calculated. grid Multi-directional average GLCM entropy E for each grid avg (x,y), including:

[0025] The meshed image grid Each grid cell is quantized to multiple gray levels;

[0026] The gray-level co-occurrence matrix of each grid in multiple directions is calculated based on multiple gray levels;

[0027] The gray-level co-occurrence matrix in each direction is normalized to a probability distribution, and the GLCM entropy in a single direction is calculated.

[0028] The average GLCM entropy in multiple directions is calculated to obtain the multi-directional average GLCM entropy E. avg (x,y), where the multi-directional average GLCM entropy E avg The mathematical expression for (x,y) is:

[0029]

[0030] In the formula, k is the direction index, K is the number of directions, n and m are both gray level indices, L is the number of gray levels, and p k (n,m) represents the probability of gray levels n and m occurring simultaneously in direction k.

[0031] In one embodiment of this application, the target grid is calculated. target The Hessian features of the inner pixels are used to locate the crack region based on the Hessian features, including:

[0032] For the target grid target Given (i,j), calculate the Hessian matrix H(i,j) for each pixel (i,j), where the mathematical expression for the Hessian matrix H(i,j) is:

[0033]

[0034] In the formula, L ii Let L be the second-order partial derivative in the i-direction. jj Let L be the second-order partial derivative in the j-direction. ij It is a mixed second-order partial derivative;

[0035] Based on the Hessian matrix H(i,j), calculate the first eigenvalue λ1 and the second eigenvalue λ2 for each pixel, where the mathematical expressions for the first eigenvalue λ1 and the second eigenvalue λ2 are as follows:

[0036]

[0037] The anisotropy degree A of each pixel is calculated based on the first feature value λ1 and the second feature value λ2 of each pixel, wherein the mathematical expression of the anisotropy degree A is:

[0038]

[0039] Pixels with an anisotropy degree A greater than or equal to a preset threshold are designated as crack pixels.

[0040] Cluster the crack pixels to obtain multiple pixel clusters, and the pixel clusters with a number of pixels greater than or equal to a preset number threshold are taken as crack pixel clusters.

[0041] The region corresponding to the cluster of crack pixels is taken as the crack region.

[0042] In one embodiment of this application, extracting pit regions from the preprocessed image includes:

[0043] An opening operation is performed on the preprocessed image to obtain a smoothed background image;

[0044] The difference between the smoothed background image and the preprocessed image is calculated to obtain the difference image diff.

[0045] The regions with gray values ​​greater than a preset filtering threshold are extracted from the difference image diff to obtain the pitted regions.

[0046] In one embodiment of this application, extracting the defect features of the crack region and the defect features of the pit region includes:

[0047] Extract the area of ​​the defect region. defect The average gray value of the defect area defect and the average gray value of the background outside the defect area Area background The defective area is a cracked area or a pitted area;

[0048] Calculate the average gray value of the defect area. defect With the background average gray value Area background The ratio of gray to gray is used to obtain the defect strength Ratio, where Ratio = gray. defect / Area background ;

[0049] Based on the area of ​​the defect region defect The defect feature is constructed using the defect intensity ratio.

[0050] In one embodiment of this application, the reference defect features of the reference defect region are compared with the comparison defect features of the comparison defect region to obtain a comparison result, including:

[0051] Based on the area A of the benchmark defect region base and the area A of the aforementioned defect region comp Calculate the area change rate ΔA, where the mathematical expression for the area change rate ΔA is:

[0052]

[0053] Extract the defect intensity of the reference defect region in the comparison image. And based on the defect intensity Ratio of the reference defect region base and the defect intensity of the comparative defect region Calculate the intensity change rate ΔRatio, where the mathematical expression for the intensity change rate ΔRatio is:

[0054]

[0055] A comparison result is constructed based on the area change rate ΔA and the intensity change rate ΔRatio.

[0056] In one embodiment of this application, generating evaluation information based on the comparison results includes:

[0057] When the area change rate ΔA is greater than or equal to a preset first change rate threshold, and the intensity change rate ΔRatio is greater than or equal to a preset second change rate threshold, the asphalt pavement repair effect is determined to be good; otherwise, the asphalt pavement is determined to have repair omissions.

[0058] This application also provides a machine vision-based intelligent assessment system for asphalt pavement repair quality, including:

[0059] The acquisition module is used to acquire a reference image before road repair and a comparison image after road repair, wherein both the reference image and the comparison image contain a standard grayscale card area;

[0060] The preprocessing module is used to preprocess the reference image and the comparison image based on the standard grayscale card area to obtain a preprocessed reference image and a preprocessed comparison image, wherein the preprocessing includes grayscale conversion, filtering and brightness unification;

[0061] The feature extraction module is used to extract a reference defect region and a reference defect feature of the reference defect region from the preprocessed reference image, and to extract a comparison defect region and a comparison defect feature of the comparison defect region from the preprocessed comparison image. The reference defect region and the comparison defect region both include crack regions and pit regions, and the reference defect feature and the comparison defect feature both include area features and intensity features.

[0062] The comparison and evaluation module is used to compare the baseline defect features of the baseline defect region with the comparison defect features of the comparison defect region to obtain comparison results; and to generate evaluation information based on the comparison results.

[0063] The beneficial effects of this invention are as follows: The intelligent evaluation method and system for asphalt pavement repair quality based on machine vision of this invention acquires a baseline image of the asphalt pavement before repair and a comparative image after repair. Then, defect features are extracted from both the baseline and comparative images, specifically the defect area and intensity before repair, and the defect area and intensity after repair. The repair effect can be evaluated by comparing these defect features. The evaluation dimensions are defect area and defect intensity. This application uses a standard grayscale card for brightness unification, thereby overcoming lighting issues. Furthermore, this application extracts features from the asphalt pavement before and after repair, including defect intensity, thus enabling a more accurate evaluation of the asphalt pavement repair effect. Attached Figure Description

[0064] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0065] Figure 1 This is an application scenario diagram of the intelligent assessment method for asphalt pavement repair quality based on machine vision, as shown in one embodiment of this application.

[0066] Figure 2 This is a flowchart illustrating a machine vision-based intelligent assessment method for asphalt pavement repair quality in one embodiment of this application;

[0067] Figure 3 This is a schematic diagram of the preprocessing flow in one embodiment of this application;

[0068] Figure 4 This is a flowchart of defect feature extraction in one embodiment of this application;

[0069] Figure 5 This is a structural diagram of an intelligent assessment system for asphalt pavement repair quality based on machine vision, as shown in one embodiment of this application.

[0070] Figure 6 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0071] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0072] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the layers related to the present invention and are not drawn according to the actual number, shape and size ratio of the layers in the actual implementation. In the actual implementation, the form and number of each layer can be arbitrarily changed, and the layer layout may also be more complex.

[0073] Numerous details are explored in the following description to provide a more thorough explanation of embodiments of the invention; however, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details.

[0074] Figure 1 This is an application scenario diagram of a machine vision-based intelligent assessment method for asphalt pavement repair quality, as shown in one embodiment of this application. Figure 1As shown, in this application, camera 110 is used to photograph the defective area 120 of the asphalt pavement. During the photographing process, a standard grayscale card 130 needs to be placed next to the defective area 120. The captured image includes both the defective area 120 and the standard grayscale card 130. The captured image is transmitted to the host device 140 for intelligent evaluation. Finally, the evaluation results are sent to the administrator 150.

[0075] Figure 2 This is a flowchart illustrating a machine vision-based intelligent assessment method for asphalt pavement repair quality in one embodiment of this application, as shown below. Figure 2 As shown: The intelligent assessment method for asphalt pavement repair quality based on machine vision in this embodiment may include steps S210 to S240:

[0076] S210, acquire a reference image before road repair and a comparison image after road repair, wherein both the reference image and the comparison image contain a standard grayscale card area;

[0077] Before the repair work can begin, it is necessary to go through Figure 1 The scene shown was captured as a reference image. base And after the repairs were completed, it was passed again. Figure 1 Comparison images of the scene shown comp To accurately quantify the defect characteristics before and after repair, it is necessary to ensure that the position, height, angle, and lighting conditions are as uniform as possible during shooting. Since lighting conditions are an uncontrollable factor (affected by weather), a standard grayscale card is used for brightness standardization. Position, height, and angle are controllable factors, and therefore, a positioning module and a support frame can be used to ensure uniformity in these aspects.

[0078] S220, preprocess the reference image and the comparison image based on the standard grayscale card area to obtain a preprocessed reference image and a preprocessed comparison image, wherein the preprocessing includes grayscale conversion, filtering and brightness unification;

[0079] In this application, image preprocessing is required before feature extraction. Figure 3 This is a schematic diagram of the preprocessing flow in one embodiment of this application, as shown below. Figure 3 As shown, the preprocessing process includes:

[0080] S221, perform grayscale conversion and high-pass filtering on the image to be processed to obtain a grayscale image, wherein the image to be processed is a reference image or a comparison image;

[0081] Since subsequent crack / pit detection relies solely on luminance information (such as GLCM entropy and Hessian features), the RGB image is converted to a single-channel grayscale image to eliminate chromaticity interference, providing a unified input for subsequent texture feature calculations and avoiding noise interference from the RGB channels. Furthermore, based on the characteristics of the defect features targeted in this application (cracks are high-frequency structures (grayscale abrupt changes), and pits are mid-frequency structures), a high-pass filter is used to remove low-frequency illumination components while preserving high-frequency details (crack and pit edges) for subsequent image processing.

[0082] S222, extract the grayscale values ​​of multiple pixels in the standard grayscale card area, and calculate the average value of the grayscale values ​​of the multiple pixels to obtain the grayscale reference value gray. ref ;

[0083] Specifically, a standard grayscale card (such as a 20% grayscale card) is fixed in a position in the image, and multiple pixels (such as a 10×10 area) are extracted to calculate the average grayscale, which serves as the reference parameter for subsequent brightness correction.

[0084] S223, Obtain the true grayscale value (gray) of the standard grayscale card. real And calculate the grayscale reference value gray. ref Compared with the actual gray value gray real The ratio k, k = gray real / gray ref Based on the ratio, the grayscale image is subjected to brightness unification processing to obtain a preprocessed image, wherein the preprocessed image is a preprocessing reference image or a preprocessing comparison image, and the mathematical expression of the preprocessed image is:

[0085] preprocessing(i,j)=k×gray(i,j)

[0086] In the formula, preprocessing(i,j) represents the gray value of pixel (i,j) in the preprocessed image, and gray(i,j) represents the gray value of pixel (i,j) in the grayscale image.

[0087] Finally, using the grayscale reference value gray ref and the true value gray real Brightness uniformization processing is performed on grayscale images to resolve feature distortion caused by lighting differences, making area / intensity features comparable across scenes.

[0088] S230, extract the reference defect region and the reference defect features of the reference defect region from the preprocessed reference image, and extract the comparison defect region and the comparison defect features of the comparison defect region from the preprocessed comparison image, wherein the reference defect region and the comparison defect region both include crack regions and pit regions, and the reference defect features and the comparison defect features both include area features and intensity features;

[0089] After preprocessing is completed, feature extraction can be performed automatically. In this application, feature extraction is performed on the preprocessed reference image and the preprocessed comparison image respectively. The extracted features are the area of ​​the crack region, the area of ​​the pit region, the defect intensity of the crack region, and the defect intensity of the pit defect.

[0090] Figure 4 This is a flowchart of defect feature extraction in one embodiment of this application, such as... Figure 4 As shown, the defect feature extraction process in this application includes:

[0091] S231, the preprocessed image is subjected to gridding processing to obtain a gridded image. grid (x,y), where (x,y) represents grid coordinates, and the preprocessing image is a preprocessing reference image or a preprocessing comparison image;

[0092] Specifically, to avoid global features masking local defects, the preprocessed image is divided into an X×Y grid (e.g., 32×32), with each grid (x, y) representing a local region. This provides local texture statistical units for subsequent GLCM entropy calculations.

[0093] S232, Calculate the gridded image. grid Road surface texture features for each grid texture The road surface texture feature texture Includes grayscale mean gray(x,y) and multi-directional average GLCM entropy E avg (x,y);

[0094] Multidirectional average GLCM entropy E avg The extraction process of (x,y) specifically includes:

[0095] S2321, the gridded image grid Each grid cell is quantized into multiple gray levels;

[0096] In this embodiment, the grayscale of pixels within the grid is mapped to L=64 grayscale levels to eliminate the difference in grayscale value range, so that GLCM calculation is performed on the same scale (in conjunction with preprocessing brightness calibration).

[0097] S2322, calculates the gray-level co-occurrence matrix of each grid in multiple directions based on multiple gray levels;

[0098] S2323 normalizes the gray-level co-occurrence matrix in each direction into a probability distribution and calculates the GLCM entropy in a single direction;

[0099] S2324, calculate the average GLCM entropy in multiple directions to obtain the multi-directional average GLCM entropy E. avg (x,y), where the multi-directional average GLCM entropy E avg The mathematical expression for (x,y) is:

[0100]

[0101] In the formula, k is the direction index, K is the number of directions, n and m are both gray level indices, L is the number of gray levels, and p k (n,m) represents the probability of gray levels n and m occurring simultaneously in direction k.

[0102] Cracked or pitted areas often appear as darker colors (corresponding to higher gray levels) due to weak reflectivity. Furthermore, these areas often present large areas of grayish-black color, exhibiting lower complexity and correspondingly lower entropy values. Therefore, this application utilizes multi-directional average GLCM entropyE. avg The texture features of each mesh are described using (x, y) and the average gray value. Furthermore, the multi-directional average GLCM entropy E in this application... avg (x,y) is based on four directions: [0,1]: horizontal direction (0°), [-1,1]: diagonal direction (45°), [-1,0]: vertical direction (90°), [-1,-1]: diagonal direction (135°).

[0103] S233, take the grid that meets the target conditions as the target grid Grid. target The target conditions include a mean grayscale value greater than a preset grayscale threshold and a multi-directional average GLCM entropy E. avg (x,y) is less than or equal to a preset entropy threshold;

[0104] In this application, the multi-directional average GLCM entropy E avg Mesh grids with (x,y) less than or equal to 4.5 and a mean gray level greater than a preset gray level are considered as potential defect meshes. This application utilizes average gray level extraction and multi-directional average GLCM entropy E. avg (x,y) calculations are used to roughly locate potential defect meshes. Then, fine-grained defect feature extraction is performed, which can effectively improve detection efficiency and accuracy.

[0105] S234, Calculate the target grid. target The Hessian features of the inner pixels are used to locate the crack region based on the Hessian features, specifically including:

[0106] In this application, Hessian features are used to extract crack regions. The principle is that cracks are linear structures, and Hessian features are very effective in describing linear structures. The specific extraction process is as follows:

[0107] S2341, for the target grid target Given (i,j), calculate the Hessian matrix H(i,j) for each pixel (i,j), where the mathematical expression for the Hessian matrix H(i,j) is:

[0108]

[0109] In the formula, L ii Let L be the second-order partial derivative in the i-direction. jj Let L be the second-order partial derivative in the j-direction. ij It is a mixed second-order partial derivative;

[0110] S2342, calculate the first eigenvalue λ1 and the second eigenvalue λ2 for each pixel based on the Hessian matrix H(i,j), wherein the mathematical expressions for the first eigenvalue λ1 and the second eigenvalue λ2 are as follows:

[0111]

[0112] S2343, calculate the anisotropy degree A of each pixel based on the first feature value λ1 and the second feature value λ2 of each pixel, wherein the mathematical expression of the anisotropy degree A is:

[0113]

[0114] S2344, pixels with anisotropy A greater than or equal to a preset threshold are designated as crack pixels;

[0115] In this application, the anisotropy degree A of each pixel is constructed based on Hessian features. For pixels with anisotropy degree A greater than 0.8, it is considered that there are linear features, and thus they are regarded as crack pixels.

[0116] S2345, cluster the crack pixels to obtain multiple pixel clusters, and take the pixel clusters with a number of pixels greater than or equal to a preset number threshold as crack pixel clusters.

[0117] To eliminate discrete noise, this application further clusters the crack pixels, retaining clusters with a large number of pixels as the pixel clusters corresponding to the crack region.

[0118] S2346, the region corresponding to the crack pixel cluster is taken as the crack region.

[0119] Finally, the crack region is constructed based on the pixel clusters corresponding to the crack region. In the above process, the mesh is first initially screened using the average gray value and multi-directional average GLCM entropy to locate potential defect meshes. Then, based on the characteristics of the crack, the Hessian features of the potential defect meshes are extracted, and the anisotropy degree is calculated to extract regions with strong linear features as crack regions. The above process balances feature extraction efficiency and accuracy, enabling rapid and accurate location of crack defect regions.

[0120] S235, extracting the pit region from the preprocessed image, specifically including:

[0121] The extraction method for pit areas differs from that for cracks. Since pit areas (such as shallow potholes and tire tracks) also contain road surface textures, they are not suitable for extraction using Hessian features. In this application, the following extraction method is employed, utilizing the grayscale features of the pit area:

[0122] S2351, Perform an opening operation on the preprocessed image to obtain a smoothed background image;

[0123] First, smooth the image with structural elements (such as 3×3 circles) to eliminate small defects, preserve pitted areas (large areas), and filter out cracks (linear structures are smoothed).

[0124] S2352, calculate the difference between the smoothed background image and the preprocessed image to obtain the difference image diff;

[0125] After the difference is calculated, the gray value of the pit area is higher in the difference image diff (the difference between the pit area and the road surface after filling), thus presenting the pit area in a high gray level and avoiding the interference of cracks (cracks have low gray level in the difference image diff).

[0126] S2353, extract the region with gray value greater than the preset screening threshold from the difference image diff to obtain the pit region.

[0127] Finally, a filtering threshold is used to filter and retain high grayscale pixels, thereby locating the pit area.

[0128] S236, Extract the defect features of the crack region and the defect features of the pit region, specifically including:

[0129] S2361, Extract the area of ​​the defect region. defect The average gray value of the defect area defect and the average gray value of the background outside the defect area Area background The defective area is a cracked area or a pitted area;

[0130] S2362, Calculate the average gray value of the defective region. defect With the background average gray value Area background The ratio of gray to gray is used to obtain the defect strength Ratio, where Ratio = gray. defect / Area background ;

[0131] S2363, based on the area of ​​the defective region. defect The defect feature is constructed using the defect intensity ratio.

[0132] In this application, for the defect regions and pit regions extracted above, the area (number of pixels) and defect intensity (contrast, quantifying the visual salience of the defect) are extracted respectively to construct the defect feature (DefectFeature). This feature is then used as the output for subsequent feature comparison.

[0133] In addition, the above process is used to extract defect features for both the baseline image and the comparison image.

[0134] S240, compare the baseline defect features of the baseline defect region with the comparative defect features of the comparative defect region to obtain a comparison result; and generate evaluation information based on the comparison result. Specifically, the process of defect feature comparison and effect evaluation includes:

[0135] S241, based on the area A of the reference defect region base and the area A of the aforementioned defect region comp Calculate the area change rate ΔA, where the mathematical expression for the area change rate ΔA is:

[0136]

[0137] Among them, the area change rate reflects the degree of repair of the physical coverage area. The smaller the repaired area, the closer ΔA is to 100%, which proves that the repair is more thorough.

[0138] S242, Extract the defect intensity of the reference defect region in the comparison image. And based on the defect intensity Ratio of the reference defect region baseand the defect intensity of the comparative defect region Calculate the intensity change rate ΔRatio, where the mathematical expression for the intensity change rate ΔRatio is:

[0139]

[0140] To accurately enhance the consistency of the repair, this application uses the intensity change rate ΔRatio of the original defect area before and after repair to reflect the consistency of the grayscale effect, and compares the defect intensity of the repaired defect area. The smaller the value, the closer ΔRatio is to 100%, which proves that the repair is more thorough.

[0141] S243, construct a comparison result Result based on the area change rate ΔA and the intensity change rate ΔRatio. comp .

[0142] S244, when the area change rate ΔA is greater than or equal to a preset first change rate threshold, and the strength change rate ΔRatio is greater than or equal to a preset second change rate threshold, the asphalt pavement repair effect is determined to be good; otherwise, the asphalt pavement is determined to have repair omissions.

[0143] Finally, the repair effect is evaluated using the repair area and the number of repairs performed in a single instance. If ΔA ≥ 0.8, the repair effect is considered good, indicating that the repair area is covered and the consistency of the modification is ensured. Otherwise, it suggests that the repair may be incomplete, in which case the baseline image should be used. base Image repair comp And management personnel who conduct evaluations based on the occurrence of results.

[0144] This invention discloses a machine vision-based intelligent evaluation method for asphalt pavement repair quality. The method acquires a baseline image of the asphalt pavement before repair and a comparison image after repair. Defect features are then extracted from both images, specifically the defect area and intensity before and after repair. The repair effect can be evaluated by comparing these defect features. The evaluation dimensions are defect area and defect intensity. This invention uses a standard grayscale card for brightness unification, thus overcoming lighting issues. Furthermore, the method extracts features from both the pre- and post-repair asphalt pavement images, including defect intensity, thus enabling a more accurate evaluation of the asphalt pavement repair effect.

[0145] like Figure 5 As shown, this application also provides a machine vision-based intelligent assessment system for asphalt pavement repair quality, including:

[0146] The acquisition module is used to acquire a reference image before road repair and a comparison image after road repair, wherein both the reference image and the comparison image contain a standard grayscale card area;

[0147] The preprocessing module is used to preprocess the reference image and the comparison image based on the standard grayscale card area to obtain a preprocessed reference image and a preprocessed comparison image, wherein the preprocessing includes grayscale conversion, filtering and brightness unification;

[0148] The feature extraction module is used to extract a reference defect region and a reference defect feature of the reference defect region from the preprocessed reference image, and to extract a comparison defect region and a comparison defect feature of the comparison defect region from the preprocessed comparison image. The reference defect region and the comparison defect region both include crack regions and pit regions, and the reference defect feature and the comparison defect feature both include area features and intensity features.

[0149] The comparison and evaluation module is used to compare the baseline defect features of the baseline defect region with the comparison defect features of the comparison defect region to obtain comparison results; and to generate evaluation information based on the comparison results.

[0150] This invention discloses a machine vision-based intelligent evaluation system for asphalt pavement repair quality. The system acquires a baseline image of the asphalt pavement before repair and a comparative image after repair. Defect features are then extracted from both images, specifically the defect area and intensity before and after repair. The repair effect can be evaluated by comparing these defect features. The evaluation dimensions are defect area and defect intensity. This system uses a standard grayscale card for brightness unification, thus overcoming lighting issues. Furthermore, the system extracts features from both before and after repair, including defect intensity, allowing for a more accurate evaluation of the asphalt pavement repair effect.

[0151] Figure 6 A schematic diagram of a computer system suitable for implementing the embodiments of this application is shown. It should be noted that... Figure 6 The computer system of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0152] like Figure 6As shown, the computer system includes a Central Processing Unit (CPU) 601, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on programs stored in Read-Only Memory (ROM) 602 or programs loaded from storage portion 608 into Random Access Memory (RAM) 603. The RAM 603 also stores various programs and data required for system operation. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An Input / Output (I / O) interface 605 is also connected to the bus 604.

[0153] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 605 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed into storage section 608 as needed.

[0154] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611. When the computer program is executed by central processing unit (CPU) 601, it performs various functions defined in the system of this application.

[0155] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0156] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0157] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0158] Another aspect of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a computer's processor, causes the computer to perform the method as described above. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not assembled into the electronic device.

[0159] Another aspect of this application provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the various embodiments described above.

[0160] The above embodiments are merely preferred embodiments provided to fully illustrate this application, and the scope of protection of this application is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on this application are all within the scope of protection of this application.

Claims

1. A machine vision-based intelligent assessment method for asphalt pavement repair quality, characterized in that, Including the following steps: Acquire a reference image before road surface repair and a comparison image after road surface repair, wherein both the reference image and the comparison image contain a standard grayscale card area; The reference image and the comparison image are preprocessed based on the standard grayscale card area to obtain a preprocessed reference image and a preprocessed comparison image. The preprocessing includes grayscale conversion, filtering, and brightness unification. The reference defect region and the reference defect feature of the reference defect region are extracted from the preprocessed reference image, and the comparison defect region and the comparison defect feature of the comparison defect region are extracted from the preprocessed comparison image. The reference defect region and the comparison defect region both include crack regions and pit regions, and the reference defect feature and the comparison defect feature both include area features and intensity features. The baseline defect features of the baseline defect region are compared with the comparative defect features of the comparative defect region to obtain a comparison result; and evaluation information is generated based on the comparison result.

2. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 1, characterized in that, Based on the standard grayscale card region, the reference image and the comparison image are preprocessed to obtain a preprocessed reference image and a preprocessed comparison image, including: The image to be processed is converted to grayscale and high-pass filtered to obtain a grayscale image, wherein the image to be processed is a reference image or a comparison image; Extract the grayscale values ​​of multiple pixels in the standard grayscale card area, and calculate the average of the grayscale values ​​of the multiple pixels to obtain the grayscale reference value gray. ref ; Obtain the true grayscale value of a standard grayscale card. real And calculate the grayscale reference value gray. ref Compared with the actual gray value gray real The ratio k, k = gray real / gray ref Based on the ratio, the grayscale image is subjected to brightness unification processing to obtain a preprocessed image, wherein the preprocessed image is a preprocessing reference image or a preprocessing comparison image, and the mathematical expression of the preprocessed image is: preprocessing(i,j)=k×gray(i,j) In the formula, preprocessing(i,j) represents the gray value of pixel (i,j) in the preprocessed image, and gray(i,j) represents the gray value of pixel (i,j) in the grayscale image.

3. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 1, characterized in that, Extracting the reference defect region and its reference defect features from the preprocessed reference image, and extracting the contrast defect region and its contrast defect features from the preprocessed contrast image, including: The preprocessed image is then meshed to obtain a meshed image. grid (x,y), where (x,y) represents grid coordinates, and the preprocessing image is a preprocessing reference image or a preprocessing comparison image; Calculate the gridded image grid Road surface texture features for each grid texture The road surface texture feature texture Includes grayscale mean gray(x,y) and multi-directional average GLCM entropy E avg (x,y); Grids that meet the target conditions are used as the target grid. target The target conditions include a grayscale mean greater than a preset grayscale threshold and a multi-directional average GLCM entropy E. avg (x,y) is less than or equal to a preset entropy threshold; Calculate the target grid target The Hessian features of the inner pixels are used to locate the crack region based on the Hessian features; Extract the pitted regions from the preprocessed image; Extract the defect features of the crack region and the defect features of the pit region.

4. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 3, characterized in that, Calculate the gridded image grid Multi-directional average GLCM entropy E for each grid avg (x,y), including: The meshed image grid Each grid cell is quantized to multiple gray levels; The gray-level co-occurrence matrix of each grid in multiple directions is calculated based on multiple gray levels; The gray-level co-occurrence matrix in each direction is normalized to a probability distribution, and the GLCM entropy in a single direction is calculated. The average GLCM entropy in multiple directions is calculated to obtain the multi-directional average GLCM entropy E. avg (x,y), where the multi-directional average GLCM entropy E avg The mathematical expression for (x,y) is: In the formula, k is the direction index, K is the number of directions, n and m are both gray level indices, L is the number of gray levels, and p k (n,m) represents the probability of gray levels n and m occurring simultaneously in direction k.

5. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 3, characterized in that, Calculate the target grid target The Hessian features of the inner pixels are used to locate the crack region based on the Hessian features, including: For the target grid target Given (i,j), calculate the Hessian matrix H(i,j) for each pixel (i,j), where the mathematical expression for the Hessian matrix H(i,j) is: In the formula, L ii Let L be the second-order partial derivative in the i-direction. jj Let L be the second-order partial derivative in the j-direction. ij It is a mixed second-order partial derivative; Based on the Hessian matrix H(i,j), calculate the first eigenvalue λ1 and the second eigenvalue λ2 for each pixel, where the mathematical expressions for the first eigenvalue λ1 and the second eigenvalue λ2 are as follows: The anisotropy degree A of each pixel is calculated based on the first feature value λ1 and the second feature value λ2 of each pixel, wherein the mathematical expression of the anisotropy degree A is: Pixels with an anisotropy degree A greater than or equal to a preset threshold are designated as crack pixels. Cluster the crack pixels to obtain multiple pixel clusters, and the pixel clusters with a number of pixels greater than or equal to a preset number threshold are taken as crack pixel clusters. The region corresponding to the cluster of crack pixels is taken as the crack region.

6. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 3, characterized in that, Extracting pit regions from the preprocessed image includes: An opening operation is performed on the preprocessed image to obtain a smoothed background image; The difference between the smoothed background image and the preprocessed image is calculated to obtain the difference image diff. The regions with gray values ​​greater than a preset filtering threshold are extracted from the difference image diff to obtain the pitted regions.

7. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 3, characterized in that, Extracting the defect features of the crack region and the pit region includes: Extract the area of ​​the defect region. defect The average gray value of the defect area defect and the average gray value of the background outside the defect area Area background The defective area is a cracked area or a pitted area; Calculate the average gray value of the defect area. defect With the background average gray value Area background The ratio of gray to gray is used to obtain the defect strength Ratio, where Ratio = gray. defect / Area background ; Based on the area of ​​the defect region defect The defect feature is constructed using the defect intensity ratio.

8. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 7, characterized in that, The reference defect features of the reference defect region are compared with the comparison defect features of the comparison defect region to obtain the comparison results, including: Based on the area A of the benchmark defect region base and the area A of the aforementioned defect region comp Calculate the area change rate ΔA, where the mathematical expression for the area change rate ΔA is: Extract the defect intensity of the reference defect region in the comparison image. And based on the defect intensity Ratio of the reference defect region base and the defect intensity of the comparative defect region Calculate the intensity change rate ΔRatio, where the mathematical expression for the intensity change rate ΔRatio is: A comparison result is constructed based on the area change rate ΔA and the intensity change rate ΔRatio.

9. The intelligent assessment method for asphalt pavement repair quality based on machine vision according to claim 8, characterized in that, Evaluation information is generated based on the comparison results, including: When the area change rate ΔA is greater than or equal to a preset first change rate threshold, and the intensity change rate ΔRatio is greater than or equal to a preset second change rate threshold, the asphalt pavement repair effect is determined to be good; otherwise, the asphalt pavement is determined to have repair omissions.

10. A machine vision-based intelligent assessment system for asphalt pavement repair quality, characterized in that, include: The acquisition module is used to acquire a reference image before road repair and a comparison image after road repair, wherein both the reference image and the comparison image contain a standard grayscale card area; The preprocessing module is used to preprocess the reference image and the comparison image based on the standard grayscale card area to obtain a preprocessed reference image and a preprocessed comparison image, wherein the preprocessing includes grayscale conversion, filtering and brightness unification; The feature extraction module is used to extract a reference defect region and a reference defect feature of the reference defect region from the preprocessed reference image, and to extract a comparison defect region and a comparison defect feature of the comparison defect region from the preprocessed comparison image. The reference defect region and the comparison defect region both include crack regions and pit regions, and the reference defect feature and the comparison defect feature both include area features and intensity features. The comparison and evaluation module is used to compare the baseline defect features of the baseline defect region with the comparison defect features of the comparison defect region to obtain comparison results; and to generate evaluation information based on the comparison results.

Images