Sensitivity analysis-based asphalt paving segregation detection method, system and device

Abstract

The invention belongs to the technical field of asphalt paving segregation detection, and particularly relates to an asphalt paving segregation detection method, system and device based on sensitivity analysis, and the method comprises the following steps: S1, obtaining an image of an asphalt pavement; s2, extracting various features of the image, normalizing the features, and splicing the features to obtain one-dimensional features; s3, performing sensitivity grading and fusion processing on the one-dimensional features to obtain a fusion matrix; s4, dividing the fusion matrix of the image into a training set and a test set; s5, inputting the training set into a support vector machine to train a classifier; and S6, inputting the test set into the trained classifier to obtain the paving segregation state of the asphalt pavement. According to the method, fusion features are constructed through sensitivity analysis to serve as feature values for pavement segregation detection, and the detection precision and stability are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of asphalt paving segregation detection, and in particular relates to an asphalt paving segregation detection method, system and device based on sensitivity analysis. Background Art

[0002] For asphalt pavement, segregation is one of the important reasons leading to pavement damage. At present, the commonly used methods for asphalt pavement inspection are mainly divided into two types: non-contact inspection and contact inspection.

[0003] Non-contact testing methods primarily include visual observation, structural depth, and core sampling. The visual observation method lacks clear standards and is subject to significant subjectivity, making it suitable for larger asphalt aggregates. The structural depth method, primarily based on sand spreading, can detect asphalt paving segregation, but the testing volume is large, time-consuming, and labor-intensive. Its accuracy is significantly affected by operator input, making it suitable only for asphalt paving segregation testing within a small area. The core sampling method extracts and screens core samples, calculating the percentage of each asphalt component in the mixture and determining the extent of paving segregation. This method offers the highest accuracy, but is irreversibly destructive to the pavement, and the extraction process is labor-intensive and time-consuming.

[0004] Contact testing primarily includes grayscale imaging, laser measurement, and CT scanning. Grayscale imaging requires vertical acquisition of road surface images, directly reflecting the road surface's unevenness through grayscale values. This method is significantly affected by lighting and camera angle. Laser measurement directly acquires road surface texture information through laser rangefinder scanning, but requires high road dryness and cleanliness, and the instrument is expensive. CT scanning, using CT technology, can obtain highly accurate three-dimensional information on asphalt pavement aggregate distribution and texture, but can only be performed indoors and requires core sampling.

[0005] In summary, there are currently two major technical challenges with asphalt paving segregation detection. First, the cost of asphalt paving segregation detection is high, and the detection method's accuracy and stability are poor, making it difficult to accurately determine the asphalt paving segregation status. Second, the determination of asphalt paving segregation is based on expert knowledge and requires professional support, which is labor-intensive. Finally, existing image-based asphalt paving segregation detection features cannot effectively support online real-time analysis and are difficult to meet the needs of convenient on-site applications. Therefore, research on how to improve the online and convenient nature of asphalt paving segregation detection methods is of great significance to extending road life and protecting driving safety. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method, system and device for detecting asphalt paving segregation based on sensitivity analysis, which can improve the detection accuracy and thus meet the requirements of asphalt segregation detection during the paving process.

[0007] The present invention is implemented as follows: The present invention provides an asphalt paving segregation detection method based on sensitivity analysis, comprising the following steps: S1. Acquire an image of an asphalt road surface; S2. extracting multiple features of the image, normalizing them, and concatenating them to obtain one-dimensional features; S3, performing sensitivity classification and fusion processing on the one-dimensional features to obtain a fusion matrix; S4, dividing the image fusion matrix into a training set and a test set; S5, input the training set into the support vector machine to train the classifier; S6. Input the test set into the trained classifier to obtain the paving segregation state of the asphalt pavement.

[0008] Furthermore, the features include statistical features, edge features and texture features, and their feature values ​​are normalized to the interval [-1, 1] according to the value range.

[0009] Furthermore, the step S2 includes the following sub-steps: S201. For an infinite range of values ​​between the upper and lower limits of the feature, a hyperbolic tangent function is selected to map the value to the interval [-1, 1]. S202: If there is no infinite value range for the upper and lower limits of the feature, normalize according to the following formula; , X’ represents the normalized value, X Represents the original value, X min represents the minimum value in the data set, X max Represents the maximum value in the data set.

[0010] Furthermore, the S3 includes the following sub-steps: S301, calculating the sensitivity effect of the one-dimensional feature through Morris sensitivity analysis; S302, calculating the sensitivity of the one-dimensional feature according to the sensitivity effect; S303, selecting a one-dimensional feature whose sensitivity meets the preset conditions to construct a model; S304, calculating the main effect coefficient of each one-dimensional feature according to the Sobol sensitivity analysis and the model; S305, calculating the fusion weight corresponding to each one-dimensional feature according to the main effect coefficient; S306 : Fusing the one-dimensional features according to the fusion weights to obtain a fusion matrix.

[0011] Furthermore, the calculation formula of the sensitivity effect is: ; In the above formula, Δ represents the perturbation increment of the i-th variable, and the function f(X) represents a k The model with independent inputs, , which are defined in k dimensional input space middle; The sensitivity is calculated as follows: ; In the above formula, Indicates sensitive effect; The expression of the model is: ; In the above formula, represents the model output, represents the parameters selected by Morris sensitivity analysis; The calculation formula of the main effect coefficient is: ; In the above formula, Represents one-dimensional features The main effect of ; ; ; The expression of the fusion weight is: ; The expression of the fusion matrix is: ; Among them, i represents the number of features used for feature fusion, It is a feature The fusion weight of .

[0012] Furthermore, S5 includes the following steps: S501, marking the image into two categories: segregated asphalt pavement and normal asphalt pavement, and using 0 and 1 to represent segregated asphalt pavement and normal asphalt pavement respectively; S502 , selecting a training set and inputting it into a support vector machine for training, using a radial basis kernel function as the kernel function, a penalty coefficient of 30, and a Gamma kernel function of 0.45.

[0013] Based on the same inventive concept, an asphalt paving segregation detection system based on sensitivity analysis includes: An acquisition module, used for acquiring an image of an asphalt road surface; Normalization module, used to extract multiple features of the image, normalize them and splice them to obtain one-dimensional features; The fusion module is used to perform sensitivity classification and fusion processing on one-dimensional features to obtain a fusion matrix; A partitioning module is used to divide the image fusion matrix into a training set and a test set; A training module is used to input the training set into the support vector machine to train the classifier; The test module is used to input the test set into the trained classifier to obtain the paving segregation state of the asphalt pavement.

[0014] Based on the same inventive concept, the present invention also provides an asphalt paving segregation detection device based on sensitivity analysis, including a processor and a memory, the memory is used to store a computer program, and the processor is used to execute the computer program to implement the steps of the above method.

[0015] The beneficial effects brought by the present invention are: (1) The present invention constructs fusion features as characteristic values ​​for paving segregation detection through sensitivity analysis, thereby improving the accuracy and stability of detection.

[0016] (2) The present invention detects the paving segregation condition by using a support vector machine (SVM), and can realize paving segregation detection without the need for professionals and professional knowledge.

[0017] (3) The present invention can realize online segregation detection during the asphalt paving process, fundamentally reducing the probability of asphalt segregation events. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is a flow chart of the method of the present invention; Figure 2 This is a flow chart of S3 in the present invention; Figure 3 Partial images of the segregated asphalt pavement and the normal asphalt pavement (Figure a shows the image of the paved segregated asphalt pavement, and Figure b shows the image of the normal asphalt pavement); Figure 4 Schematic diagram of the detection results of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the scope of protection of the present invention.

[0020] Example 1: Figures 1 to 4 As shown, a method for detecting asphalt paving segregation based on sensitivity analysis includes the following steps: S1. Acquire an image of an asphalt road surface.

[0021] S2. Extract multiple features from the image and normalize them to obtain one-dimensional features. In this step, features potentially relevant to the separation are extracted from the image, including statistical features, edge features, and texture features. Their feature values ​​are normalized to the interval [-1, 1] based on the range of values. Table 1 shows the names of the extracted features and the corresponding value intervals.

[0022] Table 1 Image feature table

[0023] In this step, some features have infinite value intervals, while others do not. Therefore, normalization can be performed separately according to the characteristics of the value intervals. Specifically, S2 includes the following sub-steps: S201. For features with an infinite range of upper and lower limits, a hyperbolic tangent function is selected to map the feature to the interval [-1, 1].

[0024] S202. For features where the upper and lower limits do not have an infinite value range, normalize according to the following formula: ; X’ represents the normalized value, X Represents the original value, X min represents the minimum value in the data set, X max Represents the maximum value in the data set.

[0025] S3, perform sensitivity classification and fusion processing on the one-dimensional features to obtain a fusion matrix. This step classifies the sensitivity of the features based on Morris sensitivity analysis and fuses the features. Figure 3 As shown, specifically, S3 includes the following sub-steps: S301, calculate the sensitivity effect of one-dimensional features through Morris sensitivity analysis. In this step, assume that the function f(X) Indicates a k The model with independent inputs, , which are defined in k dimensional input space In. The value assigned to X in If the given value Only change , while all other input values ​​remain unchanged, the corresponding model output is ,in Therefore, calculate the feature X i The formula for the sensitivity effect is: ; In the above formula, Δ represents the disturbance increment of the i-th variable.

[0026] S302, calculate the sensitivity of one-dimensional features based on the sensitivity effect. This step calculates the sensitivity effect r times for each feature to study the sensitivity of X. i The sensitivity calculation formula is: ; In the above formula, Represents the sensitivity effect. Use the following formula to measure the degree of dispersion of data: ; in, σ i represents the standard deviation, μ i Indicates the i The sensitivity of the data, N Indicates the sample size.

[0027] Multiple asphalt images were selected for calculation, and the specific average calculation results are shown in Table 2: Table 2 Sensitivity of different features

[0028] S303, select a one-dimensional feature with a sensitivity that meets the preset conditions to build a model. This step selects a feature with high sensitivity to build a model, that is, According to Table 2, the mean, Sobol operator, and LBP_GLCM homogeneity features can be selected as subsequent fusion features to build the model.

[0029] S304. Calculate the main effect coefficient of each one-dimensional feature based on the Sobol sensitivity analysis and model. The model expression is: .

[0030] In the above formula, represents the model output, represents the parameter selected by Morris sensitivity analysis.

[0031] S305. Calculate the fusion weight corresponding to each one-dimensional feature based on the main effect coefficient. The calculation formula of the main effect coefficient is: ; In the above formula, Represents one-dimensional features The main effect of ; ; .

[0032] The expression of fusion weight is: .

[0033] The main effect coefficients and corresponding weights of the three selected features are calculated as shown in Table 3: Table 3 Main effect coefficients and fusion weights of different features

[0034] S306: Fusing the one-dimensional features according to the fusion weights to obtain a fusion matrix. The expression of the fusion matrix is: .

[0035] Among them, i represents the number of features used for feature fusion, It is a feature The fusion weight of .

[0036] S4. Divide the image fusion matrix into a training set and a test set.

[0037] S5. Input the training set into the support vector machine to train the classifier. S5 includes the following steps: S501, marking the image into two categories: segregated asphalt pavement and normal asphalt pavement, and using 0 and 1 to represent segregated asphalt pavement and normal asphalt pavement respectively; S502 , selecting a training set and inputting it into a support vector machine for training. The kernel function is a radial basis function (RBF) kernel function, with a penalty coefficient of 30 and a Gamma kernel function of 0.45.

[0038] S6. Extract the fusion feature matrix of the asphalt image to be detected and input it into the trained classifier to obtain the paving segregation state of the asphalt pavement. Figure 3 As shown in the figure, the present invention selects 100 images with segregation and 100 images without segregation for classification detection, and the detection results are shown in Figure 4 As shown in the figure, 197 of the 200 asphalt images were correctly classified, achieving a detection accuracy of 98.5%. Label 1 indicates that the asphalt pavement in the image has segregation, and label 0 indicates that the image is a normal asphalt pavement.

[0039] Example 2: Based on the same inventive concept, an asphalt paving segregation detection system based on sensitivity analysis includes an acquisition module, a normalization module, a fusion module, a training module and a testing module.

[0040] The acquisition module is used to acquire images of asphalt pavement.

[0041] The normalization module is used to extract multiple features of the image and normalize them to obtain one-dimensional features.

[0042] The fusion module is used to perform sensitivity grading and fusion processing on one-dimensional features to obtain a fusion matrix.

[0043] The partitioning module is used to divide the fusion matrix of the image into a training set and a test set.

[0044] The training module is used to input the training set into the support vector machine to train the classifier.

[0045] The testing module is used to input the test set into the trained classifier to obtain the paving segregation state of the asphalt pavement.

[0046] Example 3: Based on the same inventive concept, the present invention also provides an asphalt paving segregation detection device based on sensitivity analysis, including a processor and a memory, the memory is used to store computer programs, and the processor is used to execute the computer programs to implement the steps of the method in Example 1.

[0047] The above are only preferred embodiments of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting asphalt paving segregation based on sensitivity analysis, characterized in that: The steps include: S1. Acquire an image of an asphalt road surface; S2. extracting multiple features of the image, normalizing them, and concatenating them to obtain one-dimensional features; S3, performing sensitivity classification and fusion processing on the one-dimensional features to obtain a fusion matrix; S4, dividing the image fusion matrix into a training set and a test set; S5, input the training set into the support vector machine to train the classifier; S6. Input the test set into the trained classifier to obtain the paving segregation state of the asphalt pavement.

2. The asphalt paving segregation detection method based on sensitivity analysis according to claim 1, characterized in that: The features include statistical features, edge features and texture features, and their feature values ​​are normalized to the interval [-1, 1] according to the value range.

3. The asphalt paving segregation detection method based on sensitivity analysis according to claim 2, characterized in that: The S2 includes the following sub-steps: S201. For an infinite range of values ​​between the upper and lower limits of the feature, a hyperbolic tangent function is selected to map the value to the interval [-1, 1]. S202: If there is no infinite value range for the upper and lower limits of the feature, normalize according to the following formula; ， X’ represents the normalized value, X Represents the original value, X min represents the minimum value in the data set, X max Represents the maximum value in the data set.

4. The asphalt paving segregation detection method based on sensitivity analysis according to claim 1, characterized in that: The S3 includes the following sub-steps: S301, calculating the sensitivity effect of the one-dimensional feature through Morris sensitivity analysis; S302, calculating the sensitivity of the one-dimensional feature according to the sensitivity effect; S303, selecting a one-dimensional feature whose sensitivity meets the preset conditions to construct a model; S304, calculating the main effect coefficient of each one-dimensional feature according to the Sobol sensitivity analysis and the model; S305, calculating the fusion weight corresponding to each one-dimensional feature according to the main effect coefficient; S306 : Fusing the one-dimensional features according to the fusion weights to obtain a fusion matrix.

5. The asphalt paving segregation detection method based on sensitivity analysis according to claim 4, characterized in that: The calculation formula of the sensitivity effect is: ； In the above formula, Δ represents the perturbation increment of the i-th variable, and the function f(X) represents a k The model with independent inputs, , which are defined in k dimensional input space middle; The sensitivity is calculated as follows: ； In the above formula, Indicates sensitive effect; The expression of the model is: ； In the above formula, represents the model output, represents the parameters selected by Morris sensitivity analysis; The calculation formula of the main effect coefficient is: ； In the above formula, Represents one-dimensional features The main effect of ; ; ; The expression of the fusion weight is: ； The expression of the fusion matrix is: ； Among them, i represents the number of features used for feature fusion, It is a feature The fusion weight of .

6. The asphalt paving segregation detection method based on sensitivity analysis according to claim 1, characterized in that: S5 includes the following steps: S501, marking the image into two categories: segregated asphalt pavement and normal asphalt pavement, and using 0 and 1 to represent segregated asphalt pavement and normal asphalt pavement respectively; S502 , selecting a training set and inputting it into a support vector machine for training, using a radial basis kernel function as the kernel function, a penalty coefficient of 30, and a Gamma kernel function of 0.

45.

7. An asphalt paving segregation detection system based on sensitivity analysis, characterized in that: include: An acquisition module, used for acquiring an image of an asphalt road surface; A normalization module is used to extract multiple features of the image and normalize and splice them to obtain one-dimensional features; A fusion module, configured to perform sensitivity grading and fusion processing on the one-dimensional features to obtain a fusion matrix; A partitioning module is used to divide the image fusion matrix into a training set and a test set; A training module is used to input the training set into the support vector machine to train the classifier; The test module is used to input the test set into the trained classifier to obtain the paving segregation state of the asphalt pavement.

8. An asphalt paving segregation detection device based on sensitivity analysis, characterized in that: The method comprises a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to execute the computer program to implement the steps of the method according to claim 1.

Images