Intelligent road online diagnosis and evaluation system and application

Abstract

The invention discloses an intelligent road online diagnosis and evaluation system and application, and the system is characterized in that the system comprises an acquisition hardware module, a road monitoring module, an image data transmission module, a processing center module, an image data processing module, a road disease intelligent detection module, and a road evaluation module; according to the intelligent road online diagnosis and evaluation system and the application, the unmanned aerial vehicle is used for detecting, monitoring and collecting road surface image data, the collected data are transmitted in a wireless transmission mode, an image comparison processing algorithm is provided, image enhancement processing is carried out on the road surface image data, and image contour compensation is achieved; and through a feature matching analysis algorithm, features of the road surface image data are analyzed, the road disease condition is intelligently detected, and the road health condition is evaluated.

Description

Technical Field

[0001] The present invention relates to smart engineering, and in particular to a smart road online diagnosis and evaluation system and application. Background Art

[0002] With the acceleration of urbanization and rapid economic development, the construction of transportation infrastructure has received increasing attention. As an important infrastructure for social and economic development, roads have a profound impact on the sustainable development of cities, economic vitality and people's daily lives. However, with the increase in traffic flow and the surge in the number of vehicles, the carrying capacity and usage of urban roads are facing unprecedented challenges. It is well known that roads are damaged by various factors in daily use, such as traffic load, natural climate change, material aging, etc. These factors will cause cracks, potholes, settlement and other diseases on the road, seriously affecting driving safety, increasing the risk of traffic accidents, and leading to reduced transportation efficiency. Therefore, how to monitor and evaluate roads in a timely and effective manner to provide a scientific basis for maintenance and management has become an important issue facing governments and transportation management departments at all levels.

[0003] Traditional road monitoring and assessment methods rely heavily on manual inspections and regular checks, which are not only inefficient but also highly subjective, and can easily miss subtle road damage that may have serious consequences. In addition, manual inspections generally take a long time, and often fail to detect urgent road repair problems in a timely manner, causing road damage to be discovered only after it has further deteriorated, resulting in higher repair costs and longer recovery times. The existence of this series of problems has prompted road managers to urgently seek more efficient and intelligent solutions. In recent years, with the rapid development of image processing, artificial intelligence, and big data technologies, the possibility of using modern scientific and technological means to conduct automated and intelligent monitoring of roads has gradually emerged, which provides new ideas for solving the bottleneck problems in traditional road management.

[0004] In summary, with the advancement of modern science and technology, the concepts of smart transportation and smart cities have gradually gained popularity, and intelligent road monitoring and evaluation systems have emerged. It is not only an effective means to solve the drawbacks of traditional road management, but also an inevitable requirement for promoting the high-quality development of transportation infrastructure construction. With the continuous improvement of society's requirements for road safety, comfort and reliability, building an intelligent and efficient online road monitoring and evaluation system has become an indispensable part of road management. This will not only promote the refined management of urban transportation infrastructure, but also improve the overall road operation efficiency, and provide strong technical support for building a safer and more convenient traffic environment. Therefore, conducting research on the online diagnosis and evaluation system of smart roads not only has important theoretical significance, but also has broad practical application prospects. Summary of the invention

[0005] The purpose of the present invention is to provide a smart road online diagnosis and evaluation system and application, aiming to solve the problems mentioned in the above background.

[0006] In order to achieve the above-mentioned purpose, the present invention proposes a smart road online diagnosis and evaluation system and application, which is characterized in that it includes an acquisition hardware module, a road monitoring module, an image data transmission module, a processing center module, an image data processing module, a road disease intelligent detection module, and a road evaluation module; in the acquisition hardware module, an intelligent drone array is used to perform automated flight; in the road monitoring module, the road surface condition is monitored in real time through the drone array; the image data transmission module transmits the collected road surface image data through radio technology; in the processing center module, the road surface data is received, stored, and analyzed through a central server; in the image data processing module, an image comparison processing algorithm is proposed to perform image enhancement processing on the road surface image data to achieve image contour compensation; in the road disease intelligent detection module, a feature matching analysis algorithm is used to analyze the characteristics of the road surface image data and intelligently detect the road disease situation; in the road evaluation module, the health of the road surface condition is evaluated based on the output result of the road disease intelligent detection module.

[0007] Furthermore, in the acquisition hardware module, the user uses the drone array to preset the detection route for the drone array, and the drone array performs automated work based on the preset detection route.

[0008] Furthermore, the road monitoring module uses an array of drones to monitor the road surface conditions in real time and collect data on the road surface conditions.

[0009] Furthermore, the image data transmission module uses radio technology to establish a data transmission channel between the drone array and the central server to achieve the transmission and storage of road pavement image data.

[0010] Furthermore, the processing center module receives the road surface image data collected by the drone array through the central server, stores and analyzes the road surface image data, and runs the algorithm.

[0011] Furthermore, the image contrast processing algorithm enhances the road surface image data, performs filtering operations, performs overflow correction after deblurring, and realizes image contour compensation.

[0012] Furthermore, image contour compensation is performed on the road surface image data, and the detailed process is as follows:

[0013] For the collected road surface image data, the road surface image data is represented by a gray value matrix, which is expressed as follows:

[0014]

[0015] Among them, D ori Represents the gray value matrix of the original road surface image data, d 1,1 d 1,n d i,j d m,1 d m,n They respectively represent the grayscale value of the pixel in the 1st row and 1st column of the original road pavement image data, the grayscale value of the pixel in the 1st row and nth column of the original road pavement image data, the grayscale value of the pixel in the ith row and jth column of the original road pavement image data, the grayscale value of the pixel in the mth row and 1st column of the original road pavement image data, and the grayscale value of the pixel in the mth row and nth column of the original road pavement image data. Based on the grayscale value matrix of the original road pavement image data, the original road pavement image data is enhanced. The formula is as follows:

[0016]

[0017] Where d′ i,j represents the gray value of the pixel in the i-th row and j-th column of the processed original road pavement image data, min(·) represents the minimum value of the elements in the matrix, and max(·) represents the maximum value of the elements in the matrix. Then, the high-frequency part of the enhanced original road pavement image data is filtered. First, a fuzzy filter function is constructed. The function formula is as follows:

[0018]

[0019] Among them, h i,j represents the fuzzy filter function, σ i,j Represents the standard deviation of the fuzzy filter function. Based on the fuzzy filter function, high-frequency enhancement of road pavement image data is performed, and the grayscale value of the pixel is deblurred. The calculation formula is as follows:

[0020] I i,j =d′ i,j -α i,j ×h i,j ×d′ i,j

[0021] Among them, I i,j represents the gray value of the pixel in the i-th row and j-th column of the original road surface image data after deblurring, α i,j Represents the enhancement coefficient, which corrects the overflow of the pixel by the gray value of the deblurred pixel. The calculation formula is as follows:

[0022]

[0023] Among them, I′i,j represents the gray value of the pixel at the i-th row and j-th column of the original road surface image data after overflow correction, D m , D n Respectively represent the number of vertical pixels and horizontal pixels of the original road surface image data, β i,j Represents the correction coefficient, and realizes image contour compensation through overflow correction of pixel points. The image contrast processing algorithm proposed in the present invention first performs image data enhancement processing based on the grayscale value representation of the image data to improve the grayscale value contrast of the image data, and then constructs a fuzzy filter function to perform high-frequency enhancement of the road surface image data through the fuzzy filter function, and then performs overflow correction of pixel points through a deblurring operation to realize contour compensation of the image data. The image contrast processing algorithm proposed in the present invention increases the grayscale contrast of the image, realizes image contour compensation, and enhances the image edge.

[0024] Furthermore, the feature matching analysis algorithm performs image data pooling operations on the processed road pavement image data, and then reorganizes the image features, calculates the eigenvalues, and determines the road pavement damage conditions through eigenvalue matching analysis of the image data.

[0025] Furthermore, the road pavement damage condition is determined through eigenvalue matching analysis of image data. The detailed process is as follows:

[0026] For the processed road pavement image data, image features are first extracted, and then the processed road pavement image data is pooled based on the convolution kernel using the attention mechanism, as shown below:

[0027]

[0028] Among them, I hand represents the processed road surface image data, I mean represents the road surface image data after average pooling, I max represents the road pavement image data after maximum pooling, meanxpool(·) represents the average pooling operation, which takes the average value of the surrounding pixels within each pixel of the processed road pavement image data, and maxpool(·) represents the maximum pooling operation, which takes the maximum value of the surrounding pixels within each pixel of the processed road pavement image data. After the pooling operation, the road pavement image data is reorganized, and the formula is as follows:

[0029]

[0030] Among them, I mean (l,k) represents the value of the pixel in the lth row and kth column of the road surface image data after average pooling. max(l,k) represents the value of the pixel at the lth row and kth column of the road surface image data after maximum pooling, ω mean,r.l,k ,ω max,r,l,k denote the average pooling weight and the maximum pooling weight, sigmoid(·) denotes the activation function, and I r Represents the reorganized features. Based on the reorganized features, the features of the road surface image data are calculated. The formula is as follows:

[0031]

[0032] Among them, η r represents the eigenvector, γ r represents the reorganized eigenvalue vector, It represents the multiplication of the corresponding elements of the vectors. The feature vectors of the image data are compared to determine the feature similarity of the image data. The image is clustered based on the feature vectors to determine the road pavement disease type. First, different cluster feature vectors are selected as cluster centers from all image feature vectors. Clustering operations are performed based on the cluster centers. The clustering operation formula is as follows:

[0033] ρ r,i =||η r -η con,i || 2

[0034] Among them, η con,i represents the i-th cluster center, ρ r,i Denotes the eigenvector η r The cluster similarity with the i-th cluster center, ||·|| 2 Represents the 2-norm, the cluster center with the smallest feature vector and cluster similarity forms a cluster space, and then performs a clustering operation again in the cluster space, selects different cluster centers in the cluster space, and performs clustering again, and continuously iterates the clustering operation. When no new cluster space appears, the iteration ends, and the standard road pavement disease image data set is used to form a test set and a verification set, and the data of the test set is used to optimize the feature matching analysis algorithm, and the feature results of the feature matching analysis algorithm are verified using the verification set. According to the accuracy of the feature result verification of the verification set, the reliability of the feature matching analysis algorithm for intelligent detection of road pavement diseases is judged. The feature matching analysis algorithm proposed by the present invention performs maximum pooling and average pooling on the image data, and performs feature reorganization based on the maximum pooling and average pooling results, calculates the feature vector of the image data, performs clustering operations based on the feature vector, and performs matching analysis of road pavement diseases based on the clustering results. The feature matching analysis algorithm proposed by the present invention can improve the degree of feature extraction and analysis of image data.

[0035] Furthermore, the road assessment module outputs a health index based on the road pavement damage conditions analyzed by the feature matching analysis algorithm, and evaluates the urgency of road repair based on the health index.

[0036] Beneficial Effects

[0037] 1. The image contrast processing algorithm proposed in the present invention first performs image data enhancement processing based on the grayscale value representation of the image data to improve the grayscale value contrast of the image data, and then constructs a fuzzy filter function. Through the fuzzy filter function, a high-frequency enhancement operation of the road pavement image data is performed to obtain a high-frequency component, and then a deblurring operation is performed to correct the overflow of the pixel points to achieve contour compensation of the image data. The image contrast processing algorithm proposed in the present invention increases the grayscale contrast of the image, achieves contour compensation of the image, enhances the image edge, improves the image quality, and is beneficial to improving the accuracy of the algorithm operation.

[0038] 2. The feature matching analysis algorithm proposed in the present invention performs maximum pooling and average pooling on the image data, performs feature reorganization based on the maximum pooling and average pooling results, calculates the feature vector of the image data, performs clustering operations on the feature vectors based on the feature vectors, and performs matching analysis of road pavement defects based on the clustering results. The feature matching analysis algorithm proposed in the present invention can improve the degree of feature extraction and analysis of image data and improve the accuracy of judging road pavement defect conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings in the following description are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0040] Figure 1 is a system schematic diagram of the present invention; DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] In order to achieve the above-mentioned purpose, the present invention proposes a smart road online diagnosis and evaluation system and application, which is characterized in that it includes an acquisition hardware module, a road monitoring module, an image data transmission module, a processing center module, an image data processing module, a road disease intelligent detection module, and a road evaluation module; in the acquisition hardware module, an intelligent drone array is used to perform automated flight; in the road monitoring module, the road surface condition is monitored in real time through the drone array; the image data transmission module transmits the collected road surface image data through radio technology; in the processing center module, the road surface data is received, stored, and analyzed through a central server; in the image data processing module, an image comparison processing algorithm is proposed to perform image enhancement processing on the road surface image data to achieve image contour compensation; in the road disease intelligent detection module, a feature matching analysis algorithm is used to analyze the characteristics of the road surface image data and intelligently detect the road disease situation; in the road evaluation module, the health of the road surface condition is evaluated based on the output result of the road disease intelligent detection module.

[0043] Specifically, in the acquisition hardware module, the user uses the drone array to preset the detection route for the drone array, and the drone array performs automated work based on the preset detection route.

[0044] Specifically, the road monitoring module uses an array of drones to monitor road surface conditions in real time and collect data on road surface conditions.

[0045] Specifically, the image data transmission module uses radio technology to establish a data transmission channel between the drone array and the central server to achieve the transmission and storage of road pavement image data.

[0046] Specifically, the processing center module receives the road pavement image data collected by the drone array through the central server, stores and analyzes the road pavement image data, and runs the algorithm.

[0047] Specifically, the image contrast processing algorithm enhances the road surface image data, performs filtering operations, performs overflow correction after deblurring, and realizes image contour compensation. The detailed process is as follows:

[0048] For the collected road surface image data, the road surface image data is represented by a gray value matrix, which is expressed as follows:

[0049]

[0050] Among them, D ori Represents the gray value matrix of the original road surface image data, d 1,1 d 1,n d i,j dm,1 d m,n They respectively represent the grayscale value of the pixel in the 1st row and 1st column of the original road pavement image data, the grayscale value of the pixel in the 1st row and nth column of the original road pavement image data, the grayscale value of the pixel in the ith row and jth column of the original road pavement image data, the grayscale value of the pixel in the mth row and 1st column of the original road pavement image data, and the grayscale value of the pixel in the mth row and nth column of the original road pavement image data. Based on the grayscale value matrix of the original road pavement image data, the original road pavement image data is enhanced. The formula is as follows:

[0051]

[0052] Among them, d′ i,j represents the gray value of the pixel in the i-th row and j-th column of the processed original road pavement image data, min(·) represents the minimum value of the elements in the matrix, and max(·) represents the maximum value of the elements in the matrix. Then, the high-frequency part of the enhanced original road pavement image data is filtered. First, a fuzzy filter function is constructed. The function formula is as follows:

[0053]

[0054] Among them, h i,j represents the fuzzy filter function, σ i,j Represents the standard deviation of the fuzzy filter function. Based on the fuzzy filter function, high-frequency enhancement of road pavement image data is performed, and the grayscale value of the pixel is deblurred. The calculation formula is as follows:

[0055] I i,j =d′ i,j -α i,j ×h i,j ×d′ i,j

[0056] Among them, I i,j represents the gray value of the pixel in the i-th row and j-th column of the original road surface image data after deblurring, α i,j Represents the enhancement coefficient, which corrects the overflow of the pixel by the gray value of the deblurred pixel. The calculation formula is as follows:

[0057]

[0058] Among them, I′ i,j represents the gray value of the pixel at the i-th row and j-th column of the original road surface image data after overflow correction, D m , D n Respectively represent the number of vertical pixels and horizontal pixels of the original road surface image data, β i,j Represents the correction coefficient, which realizes image contour compensation through pixel overflow correction.

[0059] Specifically, the feature matching analysis algorithm performs image data pooling operation on the processed road pavement image data, and then performs image feature reorganization, calculates the eigenvalue, and judges the road pavement disease condition through the eigenvalue matching analysis of the image data. The detailed process is as follows:

[0060] For the processed road pavement image data, image features are first extracted, and then the processed road pavement image data is pooled based on the convolution kernel using the attention mechanism, as shown below:

[0061]

[0062] Among them, I hand represents the processed road surface image data, I mean represents the road surface image data after average pooling, I max represents the road pavement image data after maximum pooling, meanxpool(·) represents the average pooling operation, which takes the average value of the surrounding pixels within each pixel of the processed road pavement image data, and maxpool(·) represents the maximum pooling operation, which takes the maximum value of the surrounding pixels within each pixel of the processed road pavement image data. After the pooling operation, the road pavement image data is reorganized, and the formula is as follows:

[0063]

[0064] Among them, I mean (l,k) represents the value of the pixel in the lth row and kth column of the road surface image data after average pooling. max (l,k) represents the value of the pixel at the lth row and kth column of the road surface image data after maximum pooling, ω mean,r.l,k ,ω max,r,l,k denote the average pooling weight and the maximum pooling weight, sigmoid(·) denotes the activation function, and I r Represents the reorganized features. Based on the reorganized features, the features of the road surface image data are calculated. The formula is as follows:

[0065]

[0066] Among them, η r represents the eigenvector, γ r represents the reorganized eigenvalue vector, It represents the multiplication of the corresponding elements of the vectors. The feature vectors of the image data are compared to determine the feature similarity of the image data. The image is clustered based on the feature vectors to determine the road pavement disease type. First, different cluster feature vectors are selected as cluster centers from all image feature vectors. Clustering operations are performed based on the cluster centers. The clustering operation formula is as follows:

[0067] ρ r,i =||η r -η con,i || 2

[0068] Among them, η con,i represents the i-th cluster center, ρ r,i Denotes the eigenvector η r The cluster similarity with the i-th cluster center, ||·|| 2 It represents the 2-norm, and the cluster center with the smallest feature vector and cluster similarity forms a cluster space. The clustering operation is performed again in the cluster space. Different cluster centers are selected in the cluster space and clustering is performed again. The clustering operation is continuously iterated. When no new cluster space appears, the iteration is terminated. The standard road pavement disease image data set is used to form a test set and a verification set. The data of the test set is used to optimize the feature matching analysis algorithm. The verification set is used to verify the feature results of the feature matching analysis algorithm. According to the accuracy of the feature result verification of the verification set, the reliability of the feature matching analysis algorithm for intelligent detection of road pavement diseases is judged.

[0069] Specifically, the road assessment module analyzes the road pavement damage based on the feature matching analysis algorithm, outputs a health index, and evaluates the urgency of road repair based on the health index.

[0070] In a specific embodiment, a user monitors the road surface condition in real time through a high-definition camera installed above the road, and transmits the collected road surface image data through a wired connection; then, the road surface image data transmitted from the high-definition camera is received through a central server, and the user uses the central server to perform algorithm calculations; then, the user uses an image contrast processing algorithm to perform image enhancement processing on the road surface image data; after that, the user uses a feature matching analysis algorithm to analyze the collected road surface data and intelligently detect road damage conditions; finally, based on the output results of the road damage intelligent detection module, the health of the road surface condition is assessed to determine the urgency of the road repair.

[0071] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0072] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0073] The above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A smart road online diagnosis and evaluation system and application, characterized in that: It includes acquisition hardware module, road monitoring module, image data transmission module, processing center module, image data processing module, road disease intelligent detection module, and road assessment module. In the acquisition hardware module, an intelligent drone array is used for automated flight. In the road monitoring module, the drone array is used to monitor the road surface conditions in real time. The image data transmission module transmits the collected road pavement image data through radio technology; in the processing center module, the road pavement data is received, stored and analyzed through the central server; in the image data processing module, an image contrast processing algorithm is proposed to perform image enhancement processing on the road pavement image data to achieve image contour compensation; the road disease intelligent detection module uses a feature matching analysis algorithm to analyze the characteristics of the road pavement image data and intelligently detect road disease conditions; in the road assessment module, the health of the road pavement condition is assessed based on the output results of the road disease intelligent detection module.

2. According to claim 1, a smart road online diagnosis and evaluation system and application is characterized in that: In the acquisition hardware module, the user uses the drone array to preset the detection route for the drone array, and the drone array performs automated work based on the preset detection route.

3. The smart road online diagnosis and evaluation system and application according to claim 1 is characterized in that: The road monitoring module uses an array of drones to monitor road surface conditions in real time and collect data on road surface conditions.

4. The smart road online diagnosis and evaluation system and application according to claim 1, characterized in that: The image data transmission module uses radio technology to establish a data transmission channel between the drone array and the central server to achieve transmission and storage of road pavement image data.

5. The smart road online diagnosis and evaluation system and application according to claim 1 is characterized in that: The processing center module receives the road pavement image data collected by the drone array through the central server, stores and analyzes the road pavement image data, and runs the algorithm.

6. The smart road online diagnosis and evaluation system and application according to claim 1, characterized in that: The image contrast processing algorithm performs enhancement processing on the road surface image data, performs filtering operation, performs overflow correction after deblurring, and realizes image contour compensation.

7. The smart road online diagnosis and evaluation system and application according to claim 6, characterized in that: The detailed process of performing image contour compensation on the road surface image data is as follows: For the collected road surface image data, the road surface image data is represented by a gray value matrix, which is expressed as follows: Among them, D ori Represents the gray value matrix of the original road surface image data, d 1,1 ,d 1,n ,d i,j ,d m,1 ,d m,n They respectively represent the grayscale value of the pixel in the 1st row and 1st column of the original road pavement image data, the grayscale value of the pixel in the 1st row and nth column of the original road pavement image data, the grayscale value of the pixel in the ith row and jth column of the original road pavement image data, the grayscale value of the pixel in the mth row and 1st column of the original road pavement image data, and the grayscale value of the pixel in the mth row and nth column of the original road pavement image data. Based on the grayscale value matrix of the original road pavement image data, the original road pavement image data is enhanced. The formula is as follows: Among them, d′ i,j represents the gray value of the pixel in the i-th row and j-th column of the processed original road pavement image data, min(·) represents the minimum value of the elements in the matrix, and max(·) represents the maximum value of the elements in the matrix. Then, the high-frequency part of the enhanced original road pavement image data is filtered. First, a fuzzy filter function is constructed. The function formula is as follows: Among them, h i,j represents the fuzzy filter function, σ i,j Represents the standard deviation of the fuzzy filter function. Based on the fuzzy filter function, the high-frequency enhancement of the road surface image data is performed to obtain the high-frequency component of the image, and then the gray value of the pixel point is deblurred. The calculation formula is as follows: I i,j =d′ i,j -a i,j ×h i,j ×d′ i,j Among them, I i,j represents the gray value of the pixel in the i-th row and j-th column of the original road surface image data after deblurring, α i,j Represents the enhancement coefficient, which corrects the overflow of the pixel by the gray value of the deblurred pixel. The calculation formula is as follows: Among them, I′ i,j represents the gray value of the pixel at the i-th row and j-th column of the original road surface image data after overflow correction, D m , D n Respectively represent the number of vertical pixels and horizontal pixels of the original road surface image data, β i,j Represents the correction coefficient, which realizes image contour compensation through pixel overflow correction.

8. The smart road online diagnosis and evaluation system and application according to claim 1, characterized in that: The feature matching analysis algorithm performs an image data pooling operation on the processed road pavement image data, and then performs image feature reorganization, calculates feature values, and determines the road pavement disease condition through feature value matching analysis of the image data.

9. The smart road online diagnosis and evaluation system and application according to claim 8, characterized in that: The road pavement damage condition is determined by matching and analyzing the characteristic values ​​of the image data. The detailed process is as follows: For the processed road pavement image data, image features are first extracted, and then the processed road pavement image data is pooled based on the convolution kernel using the attention mechanism, as shown below: Among them, I hand represents the processed road surface image data, I mean represents the road surface image data after average pooling, I max represents the road pavement image data after maximum pooling, meanxpool(·) represents the average pooling operation, which takes the average value of the surrounding pixels within each pixel of the processed road pavement image data, and maxpool(·) represents the maximum pooling operation, which takes the maximum value of the surrounding pixels within each pixel of the processed road pavement image data. After the pooling operation, the road pavement image data is reorganized, and the formula is as follows: Among them, I mean (l,k) represents the value of the pixel in the lth row and kth column of the road surface image data after average pooling. max (l,k) represents the value of the pixel at the lth row and kth column of the road surface image data after maximum pooling, ω mean,r.l,k ,ω max,r,l,k denote the average pooling weight and the maximum pooling weight, sigmoid(·) denotes the activation function, and I r Represents the reorganized features. Based on the reorganized features, the features of the road surface image data are calculated. The formula is as follows: Among them, η r represents the eigenvector, γ r represents the reorganized eigenvalue vector, It represents the multiplication of the corresponding elements of the vectors. The feature vectors of the image data are compared to determine the feature similarity of the image data. The image is clustered based on the feature vectors to determine the road pavement disease type. First, different cluster feature vectors are selected as cluster centers from all image feature vectors. Clustering operations are performed based on the cluster centers. The clustering operation formula is as follows: r r,i =||h r -or con,i ||2 Among them, η con,i represents the i-th cluster center, ρ r,i Denotes the eigenvector η r The cluster similarity with the i-th cluster center, ||·||2 represents the 2-norm, the feature vector and the cluster center with the smallest cluster similarity form a cluster space, and then clustering operation is performed again in the cluster space. Different cluster centers are selected in the cluster space, and clustering is performed again. The clustering operation is continuously iterated. When no new cluster space appears, the iteration is terminated, and the standard road pavement disease image data set is used to form a test set and a verification set. The data of the test set is used to optimize the feature matching analysis algorithm, and the verification set is used to verify the feature results of the feature matching analysis algorithm. According to the accuracy of the feature result verification of the verification set, the reliability of the feature matching analysis algorithm for intelligent detection of road pavement diseases is judged.

10. The smart road online diagnosis and evaluation system and application according to claim 1, characterized in that: The road assessment module outputs a health index based on the road pavement damage conditions analyzed by the feature matching analysis algorithm, and assesses the urgency of road repair based on the health index.

Images