Highway asphalt mixture segregation detection method and system based on machine vision

Abstract

The invention discloses a highway asphalt mixture segregation detection method and system based on machine vision, and the method specifically comprises the steps: collecting a plurality of highway images, and constructing an image analysis library; preprocessing the images in the image set; performing selective smoothing processing on the preprocessed images in the image set; segmenting the image after the selective smoothing processing, and keeping the image containing particle features; after feature extraction is carried out on the image containing the particle features, various feature parameters are reserved; the machine vision technology is adopted, automatic detection of the segregation state of the expressway asphalt mixture is achieved, the characteristic parameters of the asphalt mixture are extracted, judgment is conducted through a machine learning algorithm or a deep learning model, and the segregation state can be evaluated objectively and quantitatively.

Description

Technical Field

[0001] The present invention relates to the technical field of asphalt mixture segregation detection, and in particular to a method and system for detecting asphalt mixture segregation on highways based on machine vision. Background Art

[0002] In the process of highway construction, asphalt mixture is a commonly used road material. The segregation of asphalt mixture refers to the situation that during the production, transportation and construction process, due to the density difference or shape difference between aggregates of different particle sizes, these components cannot be evenly distributed in the entire mixture, resulting in an imbalance in the proportion of components in local areas. This uneven distribution will seriously affect the performance of the road, such as durability, resistance to water damage and skid resistance, and further accelerate the aging process of the road. In order to ensure that the paved road can meet the expected design standards and maintain a good service state for a long time, it is necessary to effectively monitor and evaluate the segregation state of the asphalt mixture to timely discover and correct potential problems;

[0003] The existing asphalt mixture segregation detection method mainly relies on construction workers to observe the appearance, color, aggregate distribution and other characteristics of the mixture with the naked eye to judge the segregation state of the asphalt mixture. This manual observation method is greatly affected by human subjective factors. There are differences in the judgment standards between different people, which makes the test results subjective and inconsistent. Quantitative analysis and machine automated judgment cannot be achieved, which restricts the unmanned and intelligent development of asphalt mixture pavers. To this end, we propose a highway asphalt mixture segregation detection method and system based on machine vision. Summary of the invention

[0004] The main purpose of the present invention is to provide a method and system for detecting segregation of asphalt mixture in highways based on machine vision, which can effectively solve the problems in the background technology.

[0005] To achieve the above object, the technical solution adopted by the present invention is:

[0006] The highway asphalt mixture segregation detection method based on machine vision includes:

[0007] Collect multiple highway images, build an image analysis library, and form an image set I = {I1, I2, ..., I k}, where I i Represents the i-th element in the set;

[0008] to I i Perform grayscale processing, the grayscale processing formula is:

[0009] I gray (i,j)=0.3·I R(i,j)+0.59·I G (i,j)+0.11·I B (i,j)

[0010] Where (i, j) represents a pixel position, I R (i,j),I G (i, j) and I B (i, j) the values ​​of the red, green, and blue channels at that position;

[0011] to I gray (i, j) is denoised, and the denoising formula is:

[0012]

[0013] Where m and n represent the offset in the i and j directions respectively;

[0014] Perform contrast enhancement processing on I′(i,j), and the contrast enhancement processing formula is:

[0015] I″(i,j)=α·I′(i,j)+β

[0016] In the formula, if α>1, the image contrast is increased; if 0<α<1, the image contrast is reduced; if β>0, the brightness is increased; if β<0, the brightness is reduced;

[0017] Select a central pixel p1(i1,j1) in I″(i,j);

[0018] Set a threshold T;

[0019] Selective smoothing is performed on I″(i,j). The selective smoothing formula is:

[0020]

[0021] Where: p1(i1+m1,j1+n1) is the value of the surrounding pixels of the central pixel p1(i1,j1) selected by I″(i,j); Z is the number of surrounding pixels that meet the conditions; S is a smooth function; when p1(i1,j1)-p1(i1+m1,j1+n1)<T, S=1; otherwise, S=0;

[0022] The watershed algorithm is introduced to segment I″′(i,j);

[0023] Use the sober gradient operator to calculate the gradient of I″′(i,j) to form a gradient image;

[0024] In the gradient image, find points with lower gradient values ​​and mark them as water sinks;

[0025] Starting from each catchment point, use a flood fill algorithm;

[0026] Determine the watershed through flood filling algorithm;

[0027] Obtain the image data after segmentation by I″′(i,j) to obtain an image database θ;

[0028] The images in the image database θ containing particle features are grouped into a set Α={Α1, Α2, Α3,…,Α k}(Α k represents the kth picture in the collection);

[0029] For an image containing particle features i (Α i ∈Α) is calculated, and the calculation formula of the perimeter C is:

[0030]

[0031] In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment;

[0032] For an image containing particle features i (Α i ∈Α) is calculated, and the area M calculation formula is:

[0033]

[0034] In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment;

[0035] For an image containing particle features i (Α i ∈Α), where the focal length F of the elliptical shape feature is calculated as follows:

[0036]

[0037] Where (x1, y1) and (x2, y2) are the two endpoints of the major axis of the ellipse, and (x3, y3) and (x4, y4) are the two endpoints of the minor axis of the ellipse;

[0038] Retain Alpha i (Α i ∈Α) are the numerical parameters of C, Μ, and F;

[0039] The image containing the particle features i (Α i∈Α) the numerical parameters of C, Μ, F are integrated into a feature vector V, V = (C, M, F);

[0040] The scoring function f(V)=w·V+λ is introduced to analyze the feature vector V, where w is the weight vector, indicating the importance of each feature, and λ is the bias term, which is used to adjust the value of f(V);

[0041] If f(V)>0, the asphalt mixture segregation state is normal;

[0042] If f(V) < 0, there is a problem with the segregation state of the asphalt mixture;

[0043] In the set A of images containing particle features, find the scoring function f(V) whose calculation result is greater than zero;

[0044] Sum the values ​​of the scoring function f(V) whose calculation results are greater than zero to obtain a specific value S sum , S sum The calculation formula is:

[0045]

[0046] In the formula, 1 f(V)＞0 is an indicator function. When f(V)>0, the value of the function is 1; otherwise, it is 0.

[0047] For the image set I = {I1, I2, ..., I k} elements are re-selected;

[0048] Let the set I={I1,I2,…,I k} each element in S sum Generate data reports based on the combination of values.

[0049] The highway asphalt mixture segregation detection system based on machine vision includes an image acquisition module, an image preprocessing module, an image smoothing processing module, an image feature extraction module, a feature parameter analysis and integration module, a scoring function calculation and summation module, and a data report generation module;

[0050] The image acquisition module is used to acquire multiple highway images, build an image analysis library, and form an image collection;

[0051] The image preprocessing module is used to process the image data in the image set;

[0052] The image smoothing processing module is used to perform selective smoothing processing on the pre-processed image;

[0053] The image feature extraction module is used to extract feature parameters of the image containing particle features;

[0054] The characteristic parameter analysis and integration module is used to integrate the characteristic parameters into a characteristic vector;

[0055] The scoring function calculation and summation module is used to find and sum the scoring functions whose calculation results are greater than zero in the set of images containing particle features;

[0056] The data report generating module is used to generate a data report;

[0057] The system further includes a storage device, a processing computing device, and a computing program stored in the storage device and executable on the processing computing device.

[0058] The present invention has the following beneficial effects:

[0059] Compared with the prior art, the present invention adopts machine vision technology to realize the automated detection of the segregation state of asphalt mixture on highways. By extracting the characteristic parameters of the asphalt mixture and making judgments using a machine learning algorithm or a deep learning model, the segregation state can be objectively and quantitatively evaluated, reducing the interference of human factors and improving the accuracy and reliability of the detection results. Compared with the traditional manual observation method, the speed and efficiency of detection are greatly improved, and the generated detection report is detailed and accurate, which provides strong technical support for highway construction and maintenance, helps construction personnel make more scientific and reasonable decisions, provides technical support for the unmanned and intelligent development of asphalt mixture pavers, and promotes technological progress in the field of road construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 It is a flow chart of the method and system for detecting the segregation of asphalt mixture of highway based on machine vision of the present invention;

[0061] Figure 2 The present invention is a structural block diagram of the highway asphalt mixture segregation detection system based on machine vision. DETAILED DESCRIPTION

[0062] The present invention will be further described below in conjunction with specific implementation methods, wherein the accompanying drawings are only used for exemplary descriptions and represent only schematic diagrams rather than actual drawings, and should not be understood as limiting the present invention. In order to better illustrate the specific implementation methods of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product.

[0063] The specific implementation process of the technical solution of the present invention includes the following steps:

[0064] Step 1: Collect multiple highway images, build an image analysis library, and form an image set I = {I1, I2, …, I k}, where I iRepresents the i-th element in the set;

[0065] Step 2: Preprocess the images in the image set I;

[0066] Step 21: First draw I i , to I i Grayscale processing is performed so that the image only contains brightness information but not color information, thereby reducing the amount of image data, simplifying the calculation amount of subsequent image processing, and improving the processing speed. The grayscale image is recorded as I gray (i, j), the grayscale processing formula is:

[0067] I gray (i,j)=0.3·I R (i,j)+0.59·I G (i,j)+0.11·I B (i,j)

[0068] Where (i, j) represents a pixel position, I R (i,j),I G (i, j) and I B (i, j) the values ​​of the red, green, and blue channels at that position;

[0069] Step 22: I gray (i,j) is subjected to denoising, by calculating image I gray The pixel selected in (i, j) (the pixel is far away from the edge of the image and is located in the center of the image as much as possible) and the average value of the 8 pixels around it are used to reduce the influence of random noise and improve the image quality. The denoised image is recorded as I′(i, j). The denoising processing formula is:

[0070]

[0071] Where m and n represent the offset in the i and j directions respectively;

[0072] Step 23: Perform contrast enhancement processing on I′(i,j). By adjusting the values ​​of contrast factor α and brightness factor β, the contrast and brightness of the image are controlled to improve the visual effect of the image. The image after contrast enhancement is recorded as I″(i,j). The contrast enhancement processing formula is:

[0073] I″(i,j)=α·I′(i,j)+β

[0074] In the formula, if α>1, the image contrast is increased; if 0<α<1, the image contrast is reduced; if β>0, the brightness is increased; if β<0, the brightness is reduced;

[0075] Step 3: Perform selective smoothing on the pre-processed images in the image set;

[0076] Step 31: Select a central pixel p1(i1,j1) in I″(i,j);

[0077] Step 32: Set a threshold T (in asphalt mixture segregation detection, the integrity of edge information is crucial, so the value of T should be low enough to capture most of the edge information in the image) to determine which pixels should be considered in the smoothing process. Only when the grayscale difference between the central pixel p1(i1,j1) and the surrounding pixels is less than T, these pixels will participate in the smoothing calculation. If it is not less than T, they will not participate in the smoothing calculation, thereby retaining these useful high-frequency signals. The image after selective smoothing is recorded as I″′(i,j);

[0078] Step 33: Perform selective smoothing on I″(i,j). The selective smoothing formula is:

[0079]

[0080] Where: p1(i1+m1,j1+n1) is the value of the surrounding pixels of the central pixel p1(i1,j1) selected by I″(i,j); Z is the number of surrounding pixels that meet the conditions; S is a smooth function; when p1(i1,j1)-p1(i1+m1,j1+n1)<T, S=1; otherwise, S=0;

[0081] Step 4: Segment the image after the selective smoothing process and retain the image containing the particle features;

[0082] Step 41: Introduce the watershed algorithm to segment I″′(i,j);

[0083] Step 411: using the sober gradient operator to perform gradient calculation on I″′(i,j) to form a gradient image;

[0084] Step 412: In the gradient image, find points with lower gradient values ​​and mark them as water confluence points;

[0085] Step 413: Starting from each water confluence point, use a flood fill algorithm;

[0086] Step 414: Determine the watershed line by flood filling algorithm;

[0087] Step 42: Obtain the image data after segmentation by I″′(i,j) to obtain an image database θ;

[0088] Step 43: The images containing particle features in the image database θ are grouped into a set Α={Α1, Α2, Α3,…,Α k}(Α k represents the kth picture in the collection);

[0089] Step 5: After extracting features from the image containing particle features, various feature parameters are retained;

[0090] Step 51: Image A containing particle features i (Α i ∈Α), the perimeter features are calculated, the perimeter contour is divided into multiple small straight line segments, and the coordinates of the endpoints of each straight line segment are marked as (a1, b1), (a2, b2), ..., (a n , b n ), calculate the length of the straight line segment by the coordinates of the two end points of the straight line segment, and add the length of each straight line segment. Since the contour is closed, (a n , b n ) to (a1, b1), thereby obtaining the total perimeter C of the particle. The perimeter C is calculated as follows:

[0091]

[0092] In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment;

[0093] Step 52: Image A containing particle features i (Α i ∈Α) is used to calculate the area feature, and the endpoint coordinates of the small straight line segments divided by the perimeter contour are used as the vertex coordinates of the polygon. i , b i ) and the next vertex (a i+1 , b i+1 ) coordinate product, the last vertex (a n , b n ) and the coordinate product of the first vertex (a1, b1), the product of the b coordinate of each vertex and the a coordinate of the next vertex, the product of the b coordinate of the last vertex and the a coordinate of the first vertex, and add these products and take the absolute value, and divide the sum by 2 to get the area M of the polygon. The area M calculation formula is:

[0094]

[0095] In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment;

[0096] Step 53: Image A containing particle features i (Α i ∈Α) to calculate the shape feature, and determine the vertex coordinates of the major axis and minor axis of the elliptical shape feature, wherein the focal length F of the elliptical shape feature is calculated as follows:

[0097]

[0098] Where (x1, y1) and (x2, y2) are the two endpoints of the major axis of the ellipse, and (x3, y3) and (x4, y4) are the two endpoints of the minor axis of the ellipse;

[0099] Step 54: Retain Alpha i (Α i ∈Α) are the numerical parameters of C, Μ, and F;

[0100] Step 6: Integrate and extract the various retained feature parameters to form a feature vector, and use the feature vector to score the image containing the particle features;

[0101] Step 61: Image A containing particle features i (Α i ∈Α) the numerical parameters of C, Μ, F are integrated into a feature vector V, V = (C, M, F);

[0102] Step 62: Introduce a scoring function f(V)=w·V+λ to analyze the feature vector V, where w is a weight vector (it is necessary to perform a minimum division on w and select the best weight vector w by calculating the square length of w to avoid coincidence in the selection of w), indicating the importance of each feature, and λ is a bias term used to adjust the value of f(V);

[0103] If f(V)>0, the asphalt mixture segregation state is normal;

[0104] If f(V) < 0, there is a problem with the segregation state of the asphalt mixture;

[0105] Step 7: Sum the scores of images containing particle features that are scored as positive numbers;

[0106] Step 71: Find a scoring function f(V) whose calculation result is greater than zero in a set A of images containing particle features;

[0107] Step 72: Sum the values ​​of the scoring function f(V) whose calculation results are greater than zero to obtain a specific value S sum , S sum The calculation formula is:

[0108]

[0109] In the formula, 1 f(V)＞0 is an indicator function. When f(V)>0, the value of the function is 1; otherwise, it is 0.

[0110] Step 8: Repeat steps 2 to 7 and finally generate a data report;

[0111] Step 81: For the image set I = {I1, I2, ..., I k} elements are re-selected;

[0112] Step 82: Set I = {I1, I2, ..., I k} each element in S sum Generate data reports based on the combination of values.

[0113] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for detecting the segregation of asphalt mixture on highways based on machine vision, characterized in that: The following steps are involved: S1: Collect multiple highway images, build an image analysis library, and form an image set I = {I1, I2, ..., I k }, where I i Represents the i-th element in the set; S2: preprocessing images in the image collection; S3: performing selective smoothing on the pre-processed images in the image set; S4: segmenting the image after the selective smoothing process, and retaining the image containing the particle features; S5: After extracting features from the image containing particle features, various feature parameters are retained; S6: integrating and extracting various retained feature parameters to form a feature vector, and scoring the image containing particle features through the feature vector; S7: summing the scores of images containing particle features that are scored as positive; S8: Repeat S2 to S7 and finally generate a data report.

2. The method for detecting segregation of asphalt mixture of highway based on machine vision according to claim 1, characterized in that , the step S2 comprises the steps of: S21: To I i Perform grayscale processing, the grayscale processing formula is: I gray (i,j)=0.3·I R (i,j)+0.59·I G (i,j)+0.11·I B (i,j) Where (i, j) represents a pixel position, I R (i,j),I G (i,j) and I B (i, j) the values ​​of the red, green, and blue channels at that position; S22: To I gray (i, j) is denoised, and the denoising formula is: Where m and n represent the offset in the i and j directions respectively; S23: Perform contrast enhancement processing on I′(i,j). The contrast enhancement processing formula is: I″(i,j)=α·I′(i,j)+β In the formula, if α>1, the image contrast is increased; if 0<α<1, the image contrast is reduced; if β>0, the brightness is increased; if β<0, the brightness is reduced.

3. The method for detecting highway asphalt mixture segregation based on machine vision according to claim 1, characterized in that: The step S3 comprises the steps of: S31: Select a central pixel p1(i1,j1) in I″(i,j); S32: setting a threshold T; S33: Perform selective smoothing on I″(i,j), and the selective smoothing formula is: Where: p1(i1+m1,j1+n1) is the value of the surrounding pixels of the central pixel p1(i1,j1) selected by I″(i,j); Z is the number of surrounding pixels that meet the conditions; S is a smooth function; when p1(i1,j1)-p1(i1+m1,j1+n1)<T, S=1; Otherwise, S=0.

4. The method for detecting segregation of asphalt mixture of highway based on machine vision according to claim 1, characterized in that: The step S4 comprises the steps of: S41: Introduce the watershed algorithm to segment I″′(i,j); S42: Obtain the image data after segmentation by I″′(i,j) to obtain an image database θ; S43: The images in the image database θ containing particle features are grouped into a set A={A1, A2, A3, ..., A k }(Α k represents the kth image in the collection).

5. The method for detecting highway asphalt mixture segregation based on machine vision according to claim 4 is characterized in that: The step S41 comprises the steps of: S411: using the sober gradient operator to perform gradient calculation on I″′(i,j) to form a gradient image; S412: In the gradient image, find points with lower gradient values ​​and mark them as water confluence points; S413: Starting from each watershed, use the flood fill algorithm; S414: Determine the watershed line through a flood filling algorithm.

6. The method for detecting segregation of asphalt mixture of highway based on machine vision according to claim 1, characterized in that: The step S5 comprises the steps of: S51: Image A containing particle features i (Α i ∈Α) is calculated, and the calculation formula of the perimeter C is: In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment; S52: Image A containing particle features i (Α i ∈Α) is calculated, and the area M calculation formula is: In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment; S53: Image A containing particle features i (Α i ∈Α), where the focal length F of the elliptical shape feature is calculated as follows: Where (x1, y1) and (x2, y2) are the two endpoints of the major axis of the ellipse, and (x3, y3) and (x4, y4) are the two endpoints of the minor axis of the ellipse; S54: Retain Alpha i (Α i ∈Α) are the numerical parameters of C, M, and F.

7. The method and system for detecting highway asphalt mixture segregation based on machine vision according to claim 1, characterized in that: The step S6 comprises the steps of: S61: Image A containing particle features i (Α i ∈Α) the numerical parameters of C, Μ, F are integrated into a feature vector V, V = (C, M, F); S62: introducing a scoring function f(V)=w·V+λ to analyze the feature vector V, where w is a weight vector indicating the importance of each feature, and λ is a bias term used to adjust the value of f(V); If f(V)>0, the asphalt mixture segregation state is normal; If f(V)<0, there is a problem with the segregation state of the asphalt mixture.

8. The method for detecting highway asphalt mixture segregation based on machine vision according to claim 1, characterized in that: The step S7 comprises the steps of: S71: Find a scoring function f(V) whose calculation result is greater than zero in a set A of images containing particle features; S72: summing the values ​​of the scoring function f(V) whose calculation results are greater than zero to obtain a specific value S sum , S sum The calculation formula is: In the formula, 1 f(V)＞0 is an indicator function, when f(V)>0, the value of the function is 1; otherwise, it is 0; The step S8 comprises the steps of: S81: For the image set I = {I1, I2, ..., I k } elements are re-selected; S82: Set I = {I1, I2, ..., I k } each element in S sum Generate data reports based on the combination of values.

9. The highway asphalt mixture segregation detection system based on machine vision is characterized by: It includes image acquisition module, image preprocessing module, image smoothing processing module, image feature extraction module, feature parameter analysis and integration module, scoring function calculation and summation module, and data report generation module; The image acquisition module is used to collect multiple highway images, build an image analysis library, and form an image set I = {I1, I2, ..., I k }, where I i Represents the i-th element in the set; The image preprocessing module is used to process the image data in the image set I, including the following steps: i Perform grayscale processing, the grayscale processing formula is: I gray (i,j)=0.3·I R (i,j)+0.59·I G (i,j)+0.11·I B (i,j) Where (i, j) represents a pixel position, I R (i,j),I G (i,j) and I B (i, j) the values ​​of the red, green, and blue channels at that position; to I gray (i, j) is denoised, and the denoising formula is: Where m and n represent the offset in the i and j directions respectively; Perform contrast enhancement processing on I′(i,j), and the contrast enhancement processing formula is: I″(i,j)=α·I′(i,j)+β In the formula, if α>1, the image contrast is increased; if 0<α<1, the image contrast is reduced; if β>0, the brightness is increased; if β<0, the brightness is reduced; The image smoothing processing module is used to perform selective smoothing processing on the pre-processed image, including the following steps: selecting a central pixel p1 (i1, j1) in I″ (i, j); Set a threshold T; Selective smoothing is performed on I″(i,j). The selective smoothing formula is: Where: p1(i1+m1,j1+n1) is the value of the surrounding pixels of the central pixel p1(i1,j1) selected by I″(i,j); Z is the number of surrounding pixels that meet the conditions; S is a smooth function; when p1(i1,j1)-p1(i1+m1,j1+n1)<T, S=1; otherwise, S=0; The image feature extraction module is used to extract feature parameters of an image containing particle features, and includes the following steps: introducing a watershed algorithm to segment I″′(i,j); Use the sober gradient operator to calculate the gradient of I″′(i,j) to form a gradient image; In the gradient image, find points with lower gradient values ​​and mark them as water sinks; Starting from each catchment point, use a flood fill algorithm; Determine the watershed through flood filling algorithm; Obtain the image data after segmentation by I″′(i,j) to obtain an image database θ; The images in the image database θ containing particle features are grouped into a set Α={Α1, Α2, Α3,…,Α k }(Α k represents the kth picture in the collection); For an image containing particle features i (Α i ∈Α) is calculated, and the calculation formula of the perimeter C is: In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment; For an image containing particle features i (Α i ∈Α) is calculated, and the area M calculation formula is: In the formula, a i represents the horizontal coordinate of the i-th line segment, b i Represents the ordinate of the i-th line segment; For an image containing particle features i (Α i ∈Α), where the focal length F of the elliptical shape feature is calculated as follows: Where (x1, y1) and (x2, y2) are the two endpoints of the major axis of the ellipse, and (x3, y3) and (x4, y4) are the two endpoints of the minor axis of the ellipse; Retain Alpha i (Α i ∈Α) are the numerical parameters of C, Μ, F; The characteristic parameter analysis and integration module is used to integrate the characteristic parameters into a characteristic vector V, including the following steps: i (Α i ∈Α) the numerical parameters of C, Μ, F are integrated into a feature vector V, V = (C, M, F); The scoring function f(V)=w·V+λ is introduced to analyze the feature vector V, where w is the weight vector, indicating the importance of each feature, and λ is the bias term, which is used to adjust the value of f(V); If f(V)>0, the asphalt mixture segregation state is normal; If f(V) < 0, there is a problem with the segregation state of the asphalt mixture; The scoring function calculation and summing module is used to find and sum the scoring functions f(V) whose calculation results are greater than zero in the set A of images containing particle features, and comprises the following steps: finding the scoring function f(V) whose calculation results are greater than zero in the set A of images containing particle features; Sum the values ​​of the scoring function f(V) whose calculation results are greater than zero to obtain a specific value S sum , S sum The calculation formula is: In the formula, 1 f(V)＞0 is an indicator function, when f(V)>0, the value of the function is 1; otherwise, it is 0; The data report generation module is used to generate a data report based on the S sum Generating a data report based on the value of For the image set I = {I1, I2, ..., I k } elements are re-selected; Let the set I = {I1, I2, ..., I k } each element in S sum Generate data reports based on the combination of values.

10. The highway asphalt mixture segregation detection system based on machine vision according to claim 9, characterized in that: The system further comprises a storage device, a processing computing device and a computing program stored in the storage device and executable on the processing computing device, wherein the processing computing device can implement the steps of the method described in any one of claims 1 to 8 when executing the program.

Images