A road asphalt surface crack evaluation method based on nondestructive testing

Abstract

This invention discloses a method for evaluating cracks in road asphalt pavement based on non-destructive testing, belonging to the field of crack monitoring technology. The method includes: collecting crack data from the asphalt pavement; performing time synchronization and information fusion processing on the crack data using an image sensor to generate a high-precision crack dataset; extracting crack geometric features from the high-precision crack dataset using image processing methods; classifying and analyzing the crack geometric features using an edge detection algorithm to generate a crack analysis report; inputting the crack analysis report into a road health assessment model to analyze the impact of the cracks, assess the potential harm of cracks to the road structure, and calculate the crack impact weights; combining the crack impact weights with road health assessment indicators to obtain a crack impact assessment set; and performing data fusion and real-time updates on the crack impact assessment set to generate a crack health assessment report. This method reduces repair risks and improves repair efficiency.

Description

Technical Field

[0001] This invention relates to the field of crack monitoring technology, and in particular to a method for evaluating cracks in road asphalt pavement based on non-destructive testing. Background Technology

[0002] With the continuous development of transportation infrastructure, the maintenance and inspection of asphalt pavement has become a crucial aspect of ensuring road safety and reliability. Crack assessment methods based on non-destructive testing (NDT) technology have been widely applied in the field of road maintenance. These methods typically utilize various image sensors and sensor devices to collect real-time data on cracks in the asphalt pavement, and then combine this data with computer image processing techniques for analysis. The core advantage lies in assessing the road surface in a non-destructive manner, avoiding the physical damage to the road structure caused by traditional methods. With the development of image sensors and image processing algorithms, NDT-based road crack assessment methods are continuously being optimized, enabling real-time and efficient processing of large amounts of data, thus improving the accuracy and efficiency of road crack identification and detection.

[0003] Existing crack detection methods mostly rely on traditional sensors and manual judgment, which are easily affected by environmental factors, resulting in insufficient guarantee of data accuracy and real-time performance. Although some methods use image processing algorithms to analyze cracks, due to the complexity of crack morphology and characteristics, a single image processing method is often insufficient to meet the high-precision identification requirements of different types of cracks. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for evaluating cracks in road asphalt pavement based on non-destructive testing, which solves the problems of insufficient data accuracy and low crack identification accuracy.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for evaluating cracks in road asphalt pavement based on non-destructive testing. The method includes: collecting crack data from the asphalt pavement; performing time synchronization and information fusion processing on the crack data using an image sensor to generate a high-precision crack dataset; extracting crack geometric features from the high-precision crack dataset using image processing methods; classifying and performing depth analysis on the crack geometric features using an edge detection algorithm to generate a crack analysis report; inputting the crack analysis report into a road health assessment model to analyze the impact of the crack analysis report, assessing the potential harm of cracks to the road structure, and calculating the crack impact weights; combining the crack impact weights with road health assessment indicators to obtain a crack impact assessment set; and performing data fusion and real-time updates on the crack impact assessment set to generate a crack health assessment report.

[0007] As a preferred embodiment of the road asphalt pavement crack evaluation method based on non-destructive testing described in this invention, the specific steps of collecting crack data of the asphalt pavement and performing time synchronization and information fusion processing on the crack data using an image sensor are as follows. Crack images, strain data, and acceleration data of asphalt surface are collected to obtain asphalt surface layer crack data. High-frequency noise in the asphalt surface layer crack data is filtered out by wavelet transform method to generate a processed crack dataset. The processed crack dataset is synchronized in time using an image sensor to generate a synchronized crack dataset.

[0008] As a preferred embodiment of the road asphalt pavement crack evaluation method based on non-destructive testing described in this invention, the high-precision crack dataset is obtained by classifying the crack dataset after synchronization.

[0009] As a preferred embodiment of the non-destructive testing-based road asphalt pavement crack evaluation method of the present invention, the method employs image processing to extract crack geometric features from a high-precision crack dataset, and uses an edge detection algorithm to classify and analyze the crack geometric features to generate a crack analysis report. The specific steps are as follows: Image processing is performed on a high-precision crack dataset to obtain standard crack image data, and crack geometric features are extracted from the standard crack image data. The crack geometric features are input into a support vector machine classifier, and the crack geometric features are preprocessed and classified by support vector machine to generate a crack classification dataset. A depth analysis is performed on each crack type in the crack classification dataset to obtain crack depth parameters. Based on the crack depth parameters, the nonlinear relationship between crack depth and propagation is analyzed, and a crack analysis report is generated.

[0010] As a preferred embodiment of the non-destructive testing-based road asphalt pavement crack evaluation method of the present invention, the following steps are taken: Performing depth analysis on each crack type in the crack classification dataset to obtain crack depth parameters, and analyzing the nonlinear relationship between crack depth and propagation based on the crack depth parameters to generate a crack analysis report. Crack depth information for each crack type is extracted from the crack classification dataset, and the crack depth information is denoised to obtain crack depth parameters. The crack depth is analyzed based on the crack depth parameter to determine the trend of crack depth variation over time, and the nonlinear relationship between crack depth and propagation is evaluated to generate a crack analysis report.

[0011] As a preferred embodiment of the non-destructive testing-based road asphalt pavement crack evaluation method of the present invention, the steps of inputting the crack analysis report into the road health assessment model, analyzing the impact degree of the crack analysis report, assessing the potential harm of cracks to the road structure, and calculating the crack impact weight are as follows. A road health assessment model is established by using a multilayer perceptron neural network. The crack analysis report is input into the road health assessment model, and numerical calculations are performed on the crack analysis report. Deep features are extracted step by step through the neuron nodes of each layer to predict the impact of cracks on road health and generate road health assessment indicators. An impact analysis was conducted on road health assessment indicators to calculate the damage index between the degree of crack impact and the road structure. The damage index was then combined with the geometric characteristics of the cracks for a comprehensive assessment calculation to generate the crack impact weight.

[0012] As a preferred embodiment of the non-destructive testing-based road asphalt pavement crack evaluation method of the present invention, the following steps are taken: conducting an impact analysis on road health assessment indicators, calculating the damage index between the degree of crack impact and the road structure, and combining the damage index with crack geometric characteristics to perform a comprehensive evaluation calculation and generate crack impact weights. The impact of cracks on road structure is analyzed based on road health assessment indicators, and the impact is combined with crack depth and width data to calculate the damage index of cracks on road structure. By combining the damage index with crack geometry, the impact of crack depth and width characteristics on road structure is comprehensively evaluated, and crack impact weights are generated.

[0013] As a preferred embodiment of the road asphalt pavement crack evaluation method based on non-destructive testing described in this invention, the specific steps for combining crack impact weights with road health assessment indicators to obtain a crack impact assessment set are as follows. By combining the impact weight of cracks with road health assessment indicators, the degree of impact of cracks under different load conditions is calculated. The impact level is modeled using multi-dimensional data analysis methods to generate a crack impact assessment set.

[0014] As a preferred embodiment of the non-destructive testing-based road asphalt pavement crack evaluation method of the present invention, the specific steps for fusing and updating the crack impact assessment set in real time to generate a crack health assessment report are as follows. The crack impact assessment set is fused and updated in real time using a data fusion algorithm to obtain crack trends; The crack trends are integrated and analyzed to generate a crack health assessment report.

[0015] As a preferred embodiment of the non-destructive testing-based road asphalt pavement crack evaluation method of the present invention, the specific steps for integrating and analyzing crack trends to generate a crack health assessment report are as follows. The crack trend is classified in multiple dimensions and analyzed over time to identify the crack development trend, assess the long-term impact of the crack development trend, and generate a crack development trend report. By combining crack development trend reports with crack assessment information centralized in crack classification datasets, a comprehensive analysis of crack impact, development trends, and repair recommendations is conducted to generate a crack health assessment report.

[0016] The beneficial effects of this invention are as follows: By integrating high-precision crack data acquisition with the Internet of Things for synchronous processing, and combining image sensors to provide real-time and accurate crack data, it lays the foundation for subsequent crack analysis and health assessment; by analyzing the nonlinear relationship between crack depth and expansion through image processing and support vector machine classifier, it accurately assesses the impact of cracks on road structures and optimizes repair strategies; it effectively improves the intelligence, real-time performance, and accuracy of road health monitoring, provides a scientific basis for road maintenance decisions, reduces repair risks, and improves repair efficiency. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method for evaluating cracks in road asphalt pavement based on non-destructive testing.

[0019] Figure 2 This is a flowchart of the data acquisition and preprocessing process.

[0020] Figure 3 This is a flowchart of crack feature extraction and classification.

[0021] Figure 4 This is a flowchart for road health assessment and impact weight calculation. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 This is one embodiment of the present invention, which provides a method for evaluating cracks in road asphalt pavement based on non-destructive testing, comprising the following steps: S1. Collect crack data of asphalt pavement, and use image sensors to perform time synchronization and information fusion processing on the crack data to generate a high-precision crack dataset.

[0026] S1.1 Acquire crack images, strain data, and acceleration data of the asphalt surface to obtain asphalt surface crack data, and filter high-frequency noise in the asphalt surface crack data using wavelet transform to generate a processed crack dataset.

[0027] It should be noted that imagery, strain data, and acceleration data of the asphalt surface cracks are collected to obtain asphalt pavement crack data. The asphalt pavement crack data is acquired in real time under different environmental conditions using high-precision sensors to ensure data accuracy and comprehensiveness. Wavelet transform processing is performed on the acquired asphalt pavement crack data. The wavelet transform method, through multi-scale analysis, filters out high-frequency noise in the asphalt pavement crack data, retaining useful low-frequency signals, thereby removing the noise component from the crack data and generating a processed crack dataset.

[0028] S1.2. Synchronize the processed crack dataset with time using an image sensor to generate a synchronized crack dataset.

[0029] It should be noted that the crack images, strain data, and acceleration data of the asphalt surface acquired using image sensors contain timestamp information. Comparing the timestamps from different sensor data ensures that all data points are mapped to a uniform time scale. Through precise time calibration, the image sensor performs time alignment on the crack data, enabling precise matching of data acquired at different time points in the time dimension. A synchronization algorithm then rearranges the crack data in chronological order and eliminates data offsets caused by time differences, generating a synchronized crack dataset.

[0030] S1.3. Classify the crack dataset after synchronization to generate a high-precision crack dataset.

[0031] It should be noted that the geometric features of the cracks, including their length, width, depth, and morphology, are extracted from the synchronized crack dataset. Feature extraction algorithms are used to accurately quantify the geometric features of the cracks, and data preprocessing is performed, including noise reduction and missing data imputation, to ensure the completeness and accuracy of the feature data. A classification method is then used to analyze the crack geometric features, learning to identify and distinguish different types of cracks and classifying them based on known crack types. To ensure the accuracy of the classification results, crack features are preprocessed during the classification process to eliminate outliers and ensure the quality of the input data. After classification, the crack data is grouped according to type to generate a high-precision crack dataset.

[0032] It should also be noted that feature extraction algorithms are algorithms used to extract meaningful information from raw data. Their purpose is to transform complex raw data into features that are easy to analyze and process. For crack data, feature extraction algorithms typically process image data, sensor data, etc., to extract the geometric features of the cracks (such as length, width, depth, and morphology) and other important parameters (such as strain and acceleration).

[0033] S2. Image processing methods are used to extract crack geometric features from a high-precision crack dataset, and edge detection algorithms are used to classify and analyze the crack geometric features to generate a crack analysis report.

[0034] S2.1 Perform image processing on the high-precision crack dataset to obtain standard crack image data, and extract crack geometric features from the standard crack image data.

[0035] It should be noted that by removing noise from the image and adjusting image brightness and contrast, the crack image data is ensured to be clear and interference-free. Edge detection algorithms are used to process the crack images, extracting the crack edge information and clarifying the crack's shape and outline, thereby obtaining standard crack image data. In the standard crack image data, the geometric features of the cracks are further extracted, including the crack's length, width, depth, and morphological features. Geometric features are obtained through image segmentation, morphological processing, and feature extraction algorithms, ensuring that the geometric features of each crack are accurately quantified.

[0036] S2.2 Input the crack geometric features into the support vector machine classifier, and perform feature preprocessing and support vector machine classification on the crack geometric features to generate a crack classification dataset.

[0037] It should be noted that feature preprocessing is performed on the crack geometric features. This preprocessing includes data normalization and standardization to ensure that features at different scales are appropriately handled by the support vector machine (SVM) classifier. Outliers and noise that may exist in the crack geometric features are addressed using filtering and denoising algorithms to ensure the quality of the input data. The preprocessed crack geometric features are then input into the SVM classifier for classification. The SVM classifier distinguishes different types of cracks by constructing an optimal hyperplane. During the training phase, the SVM algorithm learns the crack geometric features and their corresponding labels in the training data to determine the optimal classification boundary. After classification, the SVM classifier groups the crack data by type and generates a crack classification dataset.

[0038] It should also be noted that the training process of a Support Vector Machine (SVM) classifier includes the following steps: using the provided training dataset, the SVM algorithm maps each data sample to a high-dimensional space. The SVM algorithm searches for a hyperplane that maximizes the separation of data points from different classes, with the nearest sample point to the hyperplane having the largest distance to it. To find the optimal separating hyperplane, the SVM algorithm adjusts the model parameters by optimizing the objective function. During training, the SVM algorithm uses a kernel function to map the data to a high-dimensional space to handle non-linearly separable problems.

[0039] S2.3 Extract crack depth information for each crack type from the crack classification dataset, and perform noise reduction on the crack depth information to obtain crack depth parameters.

[0040] It should be noted that edge detection methods are used to extract the depth values ​​of cracks at different locations. These depth values ​​are typically expressed numerically as the vertical distance from the surface to the deepest point of the crack. After extraction, the crack depth information may contain noise or errors, requiring denoising processing. Common denoising methods include using wavelet transform, Kalman filtering, or median filtering to remove irregular fluctuations and outliers from the data. The denoised crack depth information will be smoother and more accurate, generating crack depth parameters.

[0041] S2.4 Analyze the trend of crack depth over time based on crack depth parameters, evaluate the nonlinear relationship between crack depth and propagation, and generate a crack analysis report.

[0042] It should be noted that the depth values ​​in the crack depth parameter are paired with timestamps to obtain crack depth data corresponding to each time point. This process involves plotting the evolution of crack depth data over time as a curve and fitting the crack depth variation pattern over time using statistical methods. The nonlinear relationship between crack depth and propagation is then evaluated. Crack depth and propagation data will be collected, ensuring the dataset includes crack depth values ​​and corresponding crack propagation at multiple time points. An appropriate nonlinear regression model, such as an exponential model, logarithmic model, or multinomial regression, is selected to fit the data, and the model parameters are estimated using the least squares algorithm to minimize the error between the predicted values ​​of the nonlinear regression model and the actual data. The fitted regression model is then used to analyze the relationship between crack depth and propagation, revealing the influence pattern of depth changes on crack propagation. The accuracy of the nonlinear regression model is verified, ensuring it can effectively predict crack propagation trends at different depths, and a crack analysis report is generated.

[0043] It should also be noted that the crack analysis report includes data such as the crack propagation trend, the impact of the crack on the road structure, and repair recommendations. S3. Input the crack analysis report into the road health assessment model, perform an impact degree analysis on the crack analysis report, assess the potential harm of cracks to the road structure, and calculate the impact weight of cracks.

[0044] S3.1. A road health assessment model is established through a multilayer perceptron neural network. The crack analysis report is input into the road health assessment model, and numerical calculations are performed on the crack analysis report. Deep features are extracted step by step through the neuron nodes of each layer to predict the impact of cracks on road health and generate road health assessment indicators.

[0045] It should be noted that the geometric features, propagation trends, and impact weights of cracks from the crack analysis report are input into the road health assessment model. These features are processed by a neural network, which performs nonlinear transformations on the input data to progressively extract deeper features. The multilayer perceptron (MLP) neural network utilizes these deeper features for numerical calculations to predict the impact of cracks on road health. Through backpropagation, the MLP continuously adjusts its weights and biases during training, effectively learning the relationship between crack features and road health, and generating road health assessment indicators.

[0046] It should also be noted that the road health assessment model is a method based on a multilayer perceptron neural network. By inputting geometric features, expansion trends, and impact weights from a crack analysis report, it predicts the impact of cracks on the road structure. The road health assessment model progressively extracts deeper information about crack features and performs numerical calculations to assess the potential threat of cracks to road health. The predictive ability of the road health assessment model is optimized by training the neural network using a backpropagation algorithm, ultimately generating road health assessment indicators to provide a scientific basis for road repair.

[0047] The deeper characteristics include the morphological changes of cracks, their propagation rate, the dynamic changes in their depth and width, the stress distribution around the cracks, and the impact of cracks on the long-term stability of the road structure.

[0048] S3.2 Analyze the impact of cracks on the road structure based on road health assessment indicators, and combine the impact with crack depth and width data to calculate the damage index of cracks on the road structure.

[0049] It should be noted that the crack impact characteristics in road health assessment indicators are combined with crack depth and width data to quantify the geometric characteristics of the cracks, especially their depth and width. The greater the crack depth and width, the more significant the impact on the road structure. The stress distribution in the crack area is calculated, and the stress concentration in the surrounding material is analyzed. Stress concentration typically occurs on both sides of the crack, and the crack depth and width directly affect the stress distribution and magnitude. The impact of cracks on the road structure is assessed by simulating the stress transfer effect in the crack area. The potential damage of cracks to the road structure is quantified, providing a scientific basis for repair priority assessment. The impact of cracks is quantified using a damage index, which is typically calculated based on the crack's geometric characteristics, crack distribution, and its potential threat to the road structure. The crack depth, width, and impact degree are combined to generate a crack damage index for the road structure.

[0050] The expression for the stress distribution in the crack region is: ; in, The stress distribution in the crack region; This represents the crack depth. The width of the crack; The maximum stress in the crack region; The stress intensity factor is derived from the crack mechanics theory in mechanics of materials. Stress applied from the outside, such as stress caused by traffic loads and temperature changes; It is a constant, usually a mechanical property constant of the material, such as Young's modulus, used to describe the degree of deformation of the material under stress, and its role is to help calculate the effects of stress distribution and crack propagation.

[0051] S3.3 Combine the damage index with the geometric characteristics of the crack to comprehensively evaluate the impact of crack depth and width characteristics on the road structure and generate crack impact weights.

[0052] It should be noted that the potential impact of cracks on the road structure is calculated based on the crack depth and width characteristics. Different weights are assigned according to different ranges of crack depth and width, reflecting the degree of influence of crack geometry on the road structure. For example, deeper cracks may lead to greater stress concentration and should therefore be assigned a higher weight. A comprehensive evaluation method is used to combine the damage index and crack geometry characteristics to obtain a comprehensive impact value, which ultimately generates the crack impact weight.

[0053] S4. Combine the crack impact weight with road health assessment indicators to obtain a crack impact assessment set, and perform data fusion and real-time updates on the crack impact assessment set to generate a crack health assessment report.

[0054] It should be noted that existing methods typically calculate crack impact weights based on static road health assessment indicators, and the assessment results are generally not updated in real time. These methods usually rely on offline data processing and periodic inspections and assessments, lacking the ability to perform real-time data fusion and updates to reflect dynamic changes over different time periods and under different conditions. As a result, the assessment results cannot reflect the actual development of cracks in a timely manner.

[0055] This invention combines the impact weight of cracks with road health assessment indicators and uses technologies such as the Internet of Things (IoT) to synchronize and fuse data in real time, forming a dynamically updated crack impact assessment set. By updating crack health assessment data in real time and combining it with different real-time monitoring data, the assessment report becomes more accurate and timely. This method can reflect changes in cracks in real time, providing a more precise basis for health assessment and decision-making.

[0056] S4.1 Combine the impact weight of cracks with road health assessment indicators to calculate the degree of impact of cracks under different load conditions.

[0057] It should be noted that the crack impact weights should be integrated with relevant information from the road health assessment indicators to ensure a close correlation between the impact weight of each crack and the road's health condition. By combining data from the road health assessment indicators, such as pavement load conditions, crack geometry, and material properties, the impact of cracks on the road structure under different load conditions is analyzed. The crack impact weights are then combined with the stress responses under different load conditions, and the stress distribution in the crack region is analyzed through numerical simulation. Finally, by combining the stress responses under each load condition with the crack geometry, the overall impact of cracks on the road structure is determined.

[0058] S4.2. The degree of impact is modeled through multi-dimensional data analysis methods to generate a crack impact assessment set.

[0059] It should be noted that the crack impact assessment set includes various factors such as crack depth, width, development trend, material properties, and stress response under different load conditions. Regression analysis is used to construct a model of the relationships between these factors. For example, multiple linear regression or nonlinear regression methods are used to analyze the correlation between crack geometry and road damage, assessing the risk level of cracks under different conditions. Cluster analysis is used to group the data in the crack impact assessment set to identify crack types with similar characteristics, in order to better understand the risk performance of different cracks in various environments. Through these multi-dimensional data analyses, a comprehensive crack risk assessment model is established, generating the crack impact assessment set.

[0060] S4.3. The crack impact assessment set is fused and updated in real time using a data fusion algorithm to obtain crack trends.

[0061] It should be noted that crack data from different time points should be collected, including information such as crack geometry, depth, width, and crack propagation trends. Crack data from multiple sources should be merged to ensure temporal and spatial alignment of data from different sensors. During the fusion process, crack data from different time points should be synchronized to eliminate biases caused by differences in data acquisition time, ensuring that the fused crack data more accurately reflects the true changes in cracks. New crack data should be fused with historical data to keep the crack impact assessment set up-to-date. This process not only dynamically reflects crack development trends but also promptly captures changes in cracks, providing a basis for predicting future crack development trends and obtaining crack trend data.

[0062] S4.4. Perform multi-dimensional classification and time series analysis on crack trends to identify crack development trends, assess the long-term impact of crack development trends, and generate a crack development trend report.

[0063] It should be noted that time-series data of the cracks should be collected, including information on crack depth, width, and morphological changes, and the data corresponding to each time point should be recorded. Time-series analysis methods should be used to model the changes in crack depth, width, and other characteristics at different time points, extracting the periodic fluctuation characteristics of crack propagation. Fourier transform methods should be used to extract frequency characteristics during crack change, further analyzing the periodic and non-periodic fluctuation patterns of crack propagation. Regression analysis or cluster analysis methods should be used to classify crack development trends. Based on the crack propagation characteristics at different time periods, cracks should be divided into different types, and the development patterns of each type should be identified. Based on the long-term development trend of the cracks, combined with changes in crack depth, width, and morphology, the propagation risk of cracks in the future time period should be assessed, generating a crack development trend report.

[0064] S4.5 Combine the crack development trend report with the crack assessment information in the crack classification dataset to comprehensively analyze the crack impact, development trend and repair suggestions, and generate a crack health assessment report.

[0065] It should be noted that the crack propagation trend and potential impact in the crack development trend report are matched with the assessment information for each crack type in the crack classification dataset. The crack classification dataset provides the crack's geometric features, impact weights, and crack depth information. Combining the crack depth information with the trend analysis results in the crack development trend report helps to comprehensively assess the long-term impact of each type of crack on the road structure. Through a comprehensive analysis of crack impact, development trend, and repair recommendations, using statistical and optimization methods, and combining crack type and development trend, specific repair recommendations are proposed for each crack. The repair recommendations are based on the crack's current state, future propagation trend, and potential impact, ensuring the effectiveness and relevance of the repair measures, and generating a crack health assessment report.

[0066] In summary, this invention achieves the following: high-precision crack data acquisition and synchronous fusion processing via the Internet of Things, combined with image sensors to provide real-time and accurate crack data, laying the foundation for subsequent crack analysis and health assessment; through image processing and support vector machine classifier analysis of the nonlinear relationship between crack depth and extension, it accurately assesses the impact of cracks on road structures and optimizes repair strategies; effectively improving the intelligence, real-time performance, and accuracy of road health monitoring, providing a scientific basis for road maintenance decisions, reducing repair risks, and improving repair efficiency.

[0067] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for evaluating cracks in a road asphalt surface layer based on non-destructive testing, characterized by: include, Crack data of asphalt pavement are collected, and the crack data is processed by time synchronization and information fusion using an image sensor to generate a high-precision crack dataset. Image processing methods are used to extract crack geometric features from a high-precision crack dataset, and edge detection algorithms are used to classify and analyze the crack geometric features to generate a crack analysis report. The crack analysis report is input into the road health assessment model to analyze the degree of impact of the crack analysis report, assess the potential harm of cracks to the road structure, and calculate the impact weight of cracks. By combining the crack impact weight with road health assessment indicators, a crack impact assessment set is obtained. The crack impact assessment set is then fused and updated in real time to generate a crack health assessment report.

2. The non-destructive testing based method for evaluating road asphalt pavement cracks as claimed in claim 1, wherein: The process of collecting crack data from the asphalt surface layer and performing time synchronization and information fusion processing on the crack data using an image sensor is as follows: Crack images, strain data, and acceleration data of asphalt surface are collected to obtain asphalt surface layer crack data. High-frequency noise in the asphalt surface layer crack data is filtered out by wavelet transform method to generate a processed crack dataset. The processed crack dataset is synchronized in time using an image sensor to generate a synchronized crack dataset.

3. The non-destructive testing based method for evaluating road asphalt pavement cracks as claimed in claim 2, wherein: The high-precision crack dataset was obtained by classifying the crack dataset after synchronization.

4. The non-destructive testing based method for evaluating road asphalt pavement cracks as claimed in claim 3, wherein: The method employs image processing techniques to extract crack geometric features from a high-precision crack dataset, and uses edge detection algorithms to classify and perform depth analysis on these features, generating a crack analysis report. The specific steps are as follows: Image processing is performed on a high-precision crack dataset to obtain standard crack image data, and crack geometric features are extracted from the standard crack image data. The crack geometric features are input into a support vector machine classifier, and the crack geometric features are preprocessed and classified by support vector machine to generate a crack classification dataset. A depth analysis is performed on each crack type in the crack classification dataset to obtain crack depth parameters. Based on the crack depth parameters, the nonlinear relationship between crack depth and propagation is analyzed, and a crack analysis report is generated.

5. The non-destructive testing based method for evaluating road asphalt pavement cracks as claimed in claim 4, wherein: The process involves performing depth analysis on each crack type in the crack classification dataset to obtain crack depth parameters, analyzing the nonlinear relationship between crack depth and propagation based on these parameters, and generating a crack analysis report. The specific steps are as follows: Crack depth information for each crack type is extracted from the crack classification dataset, and the crack depth information is denoised to obtain crack depth parameters. The crack depth is analyzed based on the crack depth parameter to determine the trend of crack depth variation over time, and the nonlinear relationship between crack depth and propagation is evaluated to generate a crack analysis report.

6. The non-destructive testing based method of evaluating road asphalt pavement cracks as claimed in claim 5 wherein: The process involves inputting the crack analysis report into the road health assessment model, analyzing the impact of the cracks, assessing the potential harm of the cracks to the road structure, and calculating the impact weight of the cracks. The specific steps are as follows. A road health assessment model is established by using a multilayer perceptron neural network. The crack analysis report is input into the road health assessment model, and numerical calculations are performed on the crack analysis report. Deep features are extracted step by step through the neuron nodes of each layer to predict the impact of cracks on road health and generate road health assessment indicators. An impact analysis was conducted on road health assessment indicators to calculate the damage index between the degree of crack impact and the road structure. The damage index was then combined with the geometric characteristics of the cracks for a comprehensive assessment calculation to generate the crack impact weight.

7. The non-destructive testing based method of evaluating road asphalt pavement cracks as claimed in claim 6 wherein: The process involves analyzing the impact of road health assessment indicators, calculating the damage index between the degree of crack impact and the road structure, and combining the damage index with crack geometric characteristics to perform a comprehensive assessment and calculation, generating crack impact weights. The specific steps are as follows: The impact of cracks on road structure is analyzed based on road health assessment indicators, and the impact is combined with crack depth and width data to calculate the damage index of cracks on road structure. By combining the damage index with crack geometry, the impact of crack depth and width characteristics on road structure is comprehensively evaluated, and crack impact weights are generated.

8. The method for evaluating cracks in road asphalt surface layers based on non-destructive testing as described in claim 7, characterized in that: The specific steps for combining the crack impact weight with road health assessment indicators to obtain the crack impact assessment set are as follows. By combining the impact weight of cracks with road health assessment indicators, the degree of impact of cracks under different load conditions is calculated. The impact level is modeled using multi-dimensional data analysis methods to generate a crack impact assessment set.

9. The method for evaluating cracks in road asphalt surface layers based on non-destructive testing as described in claim 8, characterized in that: The specific steps for fusing and updating the crack impact assessment set in real time to generate a crack health assessment report are as follows. The crack impact assessment set is fused and updated in real time using a data fusion algorithm to obtain crack trends; The crack trends are integrated and analyzed to generate a crack health assessment report.

10. The method for evaluating cracks in road asphalt surface layers based on non-destructive testing as described in claim 9, characterized in that: The process of integrating and analyzing crack trends to generate a crack health assessment report involves the following steps: The crack trend is classified in multiple dimensions and analyzed over time to identify the crack development trend, assess the long-term impact of the crack development trend, and generate a crack development trend report. By combining crack development trend reports with crack assessment information centralized in crack classification datasets, a comprehensive analysis of crack impact, development trends, and repair recommendations is conducted to generate a crack health assessment report.

Images