Pavement structure damage detection method, device, storage medium and electronic equipment

Abstract

The present application relates to a method, device, storage medium and electronic device for detecting pavement structure damage. The method comprises: obtaining image detection data, radar detection data and sound wave detection data of a road surface to be detected; performing dual-channel identification of structural damage on the image detection data, radar detection data and sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected; determining the guidance information for using the physical model corresponding to the road surface to be detected and the setting parameters of the physical model according to the preliminary structural damage detection data; performing mechanical analysis and wave propagation analysis on the road surface to be detected respectively using each physical analysis model according to the guidance information for using the physical model and the setting parameters of the physical model, and obtaining the pavement structure damage detection data of the road surface to be detected. The adoption of this method can effectively improve the accuracy and reliability in the detection of pavement structure damage.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, storage medium and electronic device for detecting pavement structure damage. Background Art

[0002] In traditional technology, pavement structure damage detection is usually carried out by combining manual inspection with real-time sensor data monitoring. Manual inspectors conduct visual inspections on site to identify and record obvious types of damage, such as cracks, potholes, and settlements. At the same time, sensors such as accelerometers, pressure sensors, and temperature and humidity sensors are deployed in key sections or key areas to monitor the stress, deformation, and temperature changes of the pavement in real time. The dynamic information provided by the data collected by the sensors helps analyze the behavior of the pavement under different loads and environmental conditions, thereby assisting the judgment of manual inspection results.

[0003] Manual visual inspection depends on the experience and professional level of the inspectors, and there is a certain degree of subjectivity and uncertainty. Different inspectors may have different judgment criteria for road damage, and manual inspections are easily affected by factors such as time, weather, and fatigue, which may lead to missed inspections, misjudgments, or failure to detect minor damage in a timely manner. In addition, although sensors can provide real-time data, they may be affected by extreme weather (such as heavy rain, high temperature, severe cold, etc.) or harsh environments (such as strong vibration, moisture, corrosion, etc.), resulting in reduced data accuracy; resulting in a significant impact on the accuracy of traditional technologies in road structure damage detection. Summary of the invention

[0004] Based on this, it is necessary to provide a pavement structure damage detection method, device, storage medium and electronic device that can effectively improve the accuracy of pavement structure damage detection in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for detecting pavement structure damage, comprising:

[0006] Acquire image detection data, radar detection data, and sound wave detection data of the road surface to be detected;

[0007] Performing dual-channel structural damage identification on the image detection data, the radar detection data, and the sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected;

[0008] Determining, based on the preliminary structural damage detection data, the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected;

[0009] According to the guidance information for using the physical model and the parameters set for the physical model, each physical analysis model is used to perform mechanical analysis and wave propagation analysis on the road surface to be detected, so as to obtain the pavement structure damage detection data of the road surface to be detected.

[0010] In a second aspect, the present application also provides a pavement structure damage detection device, comprising:

[0011] A road surface detection data acquisition module is used to acquire image detection data, radar detection data and sound wave detection data of the road surface to be detected;

[0012] An artificial intelligence analysis module, used for performing dual-channel structural damage identification on the image detection data, the radar detection data and the sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected;

[0013] A physical parameter determination module, used to determine the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected according to the preliminary structural damage detection data;

[0014] The physical model analysis module is used to use the guidance information of the physical model and the parameters of the physical model, and use each physical analysis model to perform mechanical analysis and wave propagation analysis on the road surface to be detected, so as to obtain the road surface structure damage detection data of the road surface to be detected.

[0015] In a third aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, any step in a pavement structure damage detection method is implemented.

[0016] In a fourth aspect, the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements any step in a pavement structure damage detection method when executing the computer program.

[0017] The above-mentioned pavement structure damage detection method, device, storage medium and electronic device can fully tap the advantages of multiple sensing technologies and obtain more comprehensive and accurate pavement structure damage information by comprehensively using image detection, radar detection and sound wave detection data, combined with a dual-channel structural damage identification method. The physical model is used to guide the analysis and set parameters, and through mechanical analysis and wave propagation analysis, the essential characteristics of pavement damage are deeply explored. It can accurately identify and quantitatively evaluate different types and degrees of structural damage, overcome the limitations of a single detection method, and provide multi-dimensional damage assessment results, thereby effectively improving the accuracy and reliability of pavement structure damage detection. In addition, the introduction of physical model analysis further provides a more scientific basis for subsequent repair strategies and maintenance decisions, helps optimize resource allocation, reduces maintenance costs, and extends the service life of pavement facilities to ensure traffic safety and smoothness. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A diagram showing an application environment of a road structure damage detection method in an embodiment;

[0020] Figure 2 A schematic diagram of a process of detecting pavement structure damage in one embodiment;

[0021] Figure 3 A schematic diagram of a flow chart of a method for determining physical model usage guidance information and physical model setting parameters in one embodiment;

[0022] Figure 4 A schematic flow chart of a method for determining guidance information using a physical model in another embodiment;

[0023] Figure 5 A schematic diagram of a flow chart of a method for obtaining preliminary structural damage detection data in one embodiment;

[0024] Figure 6 A schematic diagram of a flow chart of a method for obtaining road maintenance guidance information in one embodiment;

[0025] Figure 7 is a structural block diagram of a pavement structure damage detection device in one embodiment;

[0026] Figure 8 FIG. 4 is a diagram showing the internal structure of an electronic device in one embodiment. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] The present invention provides a method for detecting road structure damage, which can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. Among them, the server 104 can be implemented with an independent server or a server cluster composed of multiple servers.

[0029] In an exemplary embodiment, Figure 2 As shown, a pavement structure damage detection method is provided, and the method is applied to Figure 1 The server in the example is used to illustrate, including the following steps 202 to 208. Among them:

[0030] Step 202, obtaining image detection data, radar detection data and sound wave detection data of the road surface to be detected.

[0031] The road surface to be inspected may be a road surface area, a section of highway or a road surface that requires damage detection and assessment, and may have structural damage such as cracks, subsidence, potholes, etc.

[0032] Among them, the image detection data can be road surface image data acquired by a camera or high-definition camera equipment, which is usually used to identify visible damage on the road surface, such as cracks, potholes, uneven surface, etc.

[0033] Among them, radar detection data can be road surface data collected by ground penetrating radar (GPR) or other radar equipment, which is mainly used to obtain depth information of the road surface and its underlying structure and identify potential deep damage or defects.

[0034] Among them, the acoustic wave detection data can be time series data obtained by the acoustic wave sensor, reflecting the propagation characteristics of acoustic waves or other waves in the road surface. These data can reveal information about deep damage to the road surface, such as the presence of cracks, the extent of crack expansion, and the overall stability of the road surface. The depth and range of damage are usually inferred through the propagation time, speed, and reflection characteristics of the acoustic wave.

[0035] Specifically, firstly, a comprehensive photograph of the road surface is taken with high-definition camera equipment, which helps to capture the detailed features of the road surface, such as cracks, potholes, peeling and other obvious damage, and provides accurate visual information to help identify surface damage. Secondly, the road surface is scanned using radar sensors (such as ground penetrating radar). Radar waves can penetrate the road surface material, which helps to detect deep structural damage, such as cracks, voids and interlayer detachment. By analyzing the reflected signals of radar waves, detailed information about the road surface structure can be provided. Finally, the acoustic wave sensor monitors the propagation of sound waves in the road surface by emitting and receiving sound wave signals. Since road surface damage usually changes the propagation speed and attenuation characteristics of sound waves, acoustic wave detection can reveal defects and damage characteristics inside the road surface.

[0036] Step 204 , performing dual-channel structural damage identification on the image detection data, the radar detection data, and the sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected.

[0037] Among them, dual-channel identification of structural damage can be a process of using two different analysis channels (such as time analysis channel and space analysis channel) to conduct joint analysis, comprehensive identification and evaluation of pavement structure damage.

[0038] The preliminary structural damage detection data may be damage data obtained by preliminary analysis of image detection data, radar detection data, and sound wave detection data, including preliminary information on surface and deep damage on the road surface.

[0039] Specifically, in the feature fusion process, independent processing channels are first set up for the data of each sensor to fully utilize the characteristics of each data. Specifically, the camera data is processed by a convolutional neural network (CNN), focusing on capturing surface damage features in the image, such as cracks and potholes; the radar data is processed by a 3D convolutional neural network (3D CNN), specifically identifying deep morphological changes in the road surface, such as the depth of cracks and structural damage; the acoustic wave data is processed by a temporal convolutional network (TCN), focusing on the wave propagation characteristics in the time series data to identify deep damage.

[0040] After each channel is processed independently, the next recognition work will be divided into two directions: time and space. For spatial feature recognition, the spatial information of camera and radar data will be integrated to capture the specific location, shape and relationship of the damage with the surrounding environment; while temporal feature recognition uses the LSTM model to process the acoustic wave data to predict the temporal evolution trend of the damage and future expansion. In the recognition process, based on the accuracy and reliability of each sensor data, a weighting mechanism is used to weight the output of spatial recognition and temporal recognition respectively, giving the acoustic wave data a higher weight to identify deep-level damage, while the camera and radar data are adjusted according to their performance in surface damage recognition. Finally, all the identified features are merged through the fully connected layer, and the system uses the Softmax classification layer to output the specific damage type (such as cracks, potholes, settlement, etc.) and its severity (such as mild, moderate, severe), and obtain preliminary structural damage detection data.

[0041] Step 206 , determining the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be inspected based on the preliminary structural damage detection data.

[0042] The physical model usage guidance information may be guidance information for determining the strategy of series, parallel or mixed execution of models and the mode of collaboration between models when executing various physical analysis models.

[0043] Among them, the physical model setting parameters can be parameters such as initial conditions, boundary conditions and material properties input in the physical analysis model. The settings of these parameters are adjusted based on preliminary detection data and guidance information to ensure that the model can truly reflect the actual road damage.

[0044] Specifically, based on the preliminary structural damage detection data, the type, location, morphology and depth of pavement damage are analyzed in detail to determine the most appropriate physical model for simulation. For example, for crack damage, if the detection data shows the presence of cracks or crack extension signs, the crack extension model is selected, which can accurately describe the crack extension process under the action of external forces; for deep structural damage, such as interlayer delamination or settlement, the pavement settlement mechanics model is selected, focusing on simulating the settlement and deformation behavior caused by loads; if the damage data shows the presence of wave propagation anomalies, the elastic wave propagation model and the waveform reflection and refraction model are used to analyze the propagation characteristics of sound waves or radar waves in the pavement to help identify potential deep damage. After selecting the appropriate model, according to the damage characteristics, depth and spatial distribution.

[0045] In order to ensure that the physical model can accurately reflect the actual damage situation, the key to the use of guidance information by the physical model lies in how to reasonably arrange the execution order and execution mode of different models. First of all, the execution strategy of the model must be determined based on the type and characteristics of the damage. For example, if the preliminary analysis shows that there are cracks and the expansion of the cracks is affected by external loads, then the crack expansion model needs to be given priority to simulate the expansion process of the cracks under the action of loads. Next, combined with the results of crack expansion, it is necessary to introduce an elastic wave propagation model to further analyze the impact of crack expansion on the propagation of sound waves or radar waves. In this case, the crack expansion model and the elastic wave propagation model will adopt a series execution strategy to analyze the changes in crack expansion and wave propagation in turn to ensure that the impact of cracks on wave propagation can be accurately captured.

[0046] On the other hand, if there is obvious settlement of the road surface, and the test data shows that the settlement is closely related to the reflection and refraction of the waveform, then the road surface settlement mechanical model and the waveform reflection and refraction model may be executed in parallel. In this case, the settlement model and the waveform model run simultaneously, respectively dealing with the impact of road surface settlement on the overall structure and the waveform propagation characteristics. Through this parallel calculation, the changes in settlement and waveform are evaluated simultaneously to ensure that the impact of the two can be analyzed synchronously. The role of the guidance information here is to help determine when to execute in series and when to execute in parallel, and to reasonably set the input and output of each model during execution to ensure that the mutual influence between different models can be fully considered.

[0047] Finally, according to the different characteristics of damage and the analysis requirements of each physical model, the guidance information helps determine the execution order and method of the model. For example, for some complex damage, it may be necessary to use a series-parallel mixed execution strategy to organically combine multiple models, considering the relationship between crack extension and wave propagation, and analyzing the impact of settlement on waveform propagation, so as to provide more accurate analysis results for damage assessment. In this way, the physical models can work together effectively to ensure that every detail in the analysis process is fully reflected.

[0048] When the physical analysis model is executed in series, parallel or mixed, the details of parameter setting depend on the dependency between models and the execution order. In series execution, the output of the previous model is used as the input of the next model, and it is necessary to ensure that the results of the previous step are accurately reflected in the subsequent model. For example, when the crack extension model is executed, the parameters such as the initial size and extension speed of the crack need to be set according to the preliminary damage data, and the results after crack extension (such as crack depth and morphology) will become the input of the elastic wave propagation model. At this time, the parameters of the elastic wave propagation model such as wave velocity and attenuation coefficient must be dynamically adjusted according to the impact of crack extension to ensure that wave propagation can truly reflect the impact of cracks on the material. In parallel execution, multiple models run simultaneously, each processing different damage characteristics. When setting parameters, it is necessary to ensure that the initial conditions of each model are independent, but consistency must be maintained in terms of shared data. For example, the settlement mechanics model and the waveform reflection model can run simultaneously. The settlement model will perform settlement analysis based on the load distribution and foundation support characteristics, while the waveform reflection model will perform wave propagation analysis based on the reflection and refraction characteristics between pavement layers. The parameters of the two models, such as the thickness of the settlement area and material properties, are set independently, but the shared pavement features such as pavement morphology and material properties need to be consistent. Hybrid execution combines the characteristics of series and parallel. The model has both the dependency of series execution and the independence of parallel execution. For example, the crack extension model and the elastic wave propagation model are executed in series. After that, the settlement mechanics model can readjust the thickness of the settlement area, support characteristics and other parameters according to the impact of crack extension on pavement settlement, to ensure that the synergistic effect of settlement and crack extension is accurately reflected. After the model parameters of each model are determined, the image detection data, radar detection data, acoustic wave detection data and preliminary structural damage detection data are input into the corresponding model, and combined with the model parameters of each previous model as the physical model setting parameters. This setting not only takes into account the independence and interdependence of each model, but also flexibly adjusts the initial conditions, boundary conditions and material properties according to the execution order to ensure the accuracy of the entire analysis process and seamless connection between models.

[0049] In a specific embodiment, when executing the crack extension model, the initial parameter settings usually include the initial size of the crack, the initial speed of crack extension, the elastic modulus of the pavement material, the fracture toughness, etc. These parameters directly determine the expansion trend of the crack during the simulation process. Assuming that the image and radar data in the previous step have provided the location and morphology of the crack, then these actual data will be used as the basic input for the initial size and extension speed of the crack to ensure that the model starts the analysis from the actual damage state. If the crack extension model completes the prediction of crack behavior, its results may need to be passed as input to the elastic wave propagation model. For example, the expansion of the crack may change the overall stiffness of the material, which will affect the propagation speed and attenuation characteristics of the sound wave or radar wave. Therefore, the initial parameters of the elastic wave propagation model (such as the density and elastic modulus of the pavement material) may need to be dynamically adjusted according to the expansion of the crack, especially in the area where the crack extends, the nonlinear characteristics of the material may need to be taken into account, thereby affecting the reflection and refraction during the wave propagation process. In the pavement settlement mechanics model, the initial parameters usually involve the thickness of the settlement area, the support characteristics of the underground layer, the load distribution of the pavement, etc. These parameters need to be adjusted based on the results of the previous crack extension and wave propagation analysis. If crack propagation leads to local settlement or deformation, the parameters of the settlement mechanics model (such as the initial thickness and material stiffness of the settlement area) must be reset in accordance with the changes caused by these cracks. In addition, the boundary conditions in the settlement mechanics model (such as the support method of the road surface, the type of external load, etc.) also need to be adjusted according to the specific road surface conditions, environmental conditions and the output results of the previous stage model to ensure the continuity and accuracy of the entire analysis process.

[0050] Step 208, according to the physical model use guidance information and physical model setting parameters, use each physical analysis model to perform mechanical analysis and wave propagation analysis on the road surface to be tested respectively, and obtain the road surface structure damage detection data of the road surface to be tested.

[0051] Among them, the pavement structure damage detection data can be the comprehensive information on the type, location, depth and severity of pavement damage obtained through methods such as mechanical analysis and wave propagation analysis.

[0052] Specifically, numerical simulation techniques (such as finite element analysis, boundary element analysis, etc.) are used to conduct detailed mechanical analysis and wave propagation analysis. In the process of mechanical analysis, a detailed mechanical model of the pavement structure is established to simulate the deformation response and stress distribution of the pavement under different loads, especially the effects of surface and deep damage such as cracks, potholes, and settlement on the mechanical properties of the pavement. Specifically, the mechanical model takes into account the elastic and plastic properties of the materials at each layer of the pavement and their interaction with the underground layer, so as to calculate the stress field and deformation field of the pavement under the action of external forces. These calculations can reveal the specific effects of damage (such as the starting point of cracks, the expansion trend, and the distribution of settlement areas) on the overall stability and bearing capacity of the pavement, and predict whether local damage may lead to structural failure or deformation in a larger range. In the wave propagation analysis, the elastic wave propagation model is used to simulate the propagation characteristics of sound waves or radar waves in the pavement, focusing on the analysis of the effects of structural damage such as cracks and settlement on the wave propagation path, propagation speed, and attenuation characteristics. By simulating the wave propagation process, the reflection, refraction, and wave velocity changes of the wave propagating in the damaged area can be accurately revealed, so as to infer the influence of cracks or settlement on the depth and range of wave propagation. This analysis method is particularly important for the detection of deep damage, because even if damage such as cracks or settlement is hidden deep, the wave propagation model can still reveal its existence and impact through the change of waves. Finally, the results of mechanical analysis and wave propagation analysis are combined to generate comprehensive pavement structure damage detection data, covering the type of damage (such as cracks, potholes, settlement, etc.), location, scale, depth and possible development trend.

[0053] In the above-mentioned pavement structure damage detection method, by comprehensively using image detection, radar detection and acoustic wave detection data, combined with a dual-channel structural damage identification method, it is possible to fully tap the advantages of multiple sensing technologies and obtain more comprehensive and accurate pavement structure damage information. By using physical models to guide analysis and set parameters, through mechanical analysis and wave propagation analysis, the essential characteristics of pavement damage are deeply explored, and different types and degrees of structural damage can be accurately identified and quantitatively evaluated. It overcomes the limitations of a single detection method and provides multi-dimensional damage assessment results, thereby effectively improving the accuracy and reliability of pavement structure damage detection. In addition, the introduction of physical model analysis further provides a more scientific basis for subsequent repair strategies and maintenance decisions, helps optimize resource allocation, reduces maintenance costs, and extends the service life of pavement facilities to ensure traffic safety and smoothness.

[0054] In an exemplary embodiment, Figure 3 As shown, according to the preliminary structural damage detection data, determining the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected includes steps 302 to 308. Among them:

[0055] Step 302: Determine at least two physical analysis models corresponding to the road surface to be inspected based on the preliminary structural damage detection data.

[0056] Among them, physical analysis models can be mathematical models that describe and simulate the process of pavement damage, deformation or failure. These models are based on physical principles such as mechanics, wave propagation, heat conduction, etc., and consider the response of the pavement under various external loads, environmental changes or internal defects (such as cracks, settlement, etc.).

[0057] Specifically, after obtaining preliminary structural damage detection data (such as crack location, morphology, depth, settlement, etc.), it is necessary to extract the key features of pavement damage from these data. Then, based on the damage type and data characteristics, select at least two appropriate physical analysis models for subsequent analysis. For example, for the detection of surface cracks or cracks, a crack propagation model can be selected to simulate the expansion behavior of cracks under external forces; if pavement settlement is involved, a pavement settlement mechanics model may be selected to simulate the compression deformation of the pavement by the load; at the same time, it can also be combined with an elastic wave propagation model or a waveform reflection and refraction model to identify deep damage through wave propagation.

[0058] Step 304: Determine analysis model serial execution information and analysis model parallel execution information according to the associations between each of the physical analysis models.

[0059] The analysis model serial execution information may be a strategy describing the execution of multiple physical analysis models in a certain order.

[0060] The analysis model parallel execution information may be a strategy for independently executing multiple physical analysis models simultaneously. In the parallel execution mode, the models are independent of each other and do not need to rely on the results of the previous model.

[0061] Specifically, after selecting the physical model, the interdependence between these models needs to be analyzed next. Some physical models may need to be executed in a certain order. For example, the crack extension model may need to rely on the results of mechanical analysis to further determine the extension range of the crack, so these models need to be executed serially. On the other hand, if there is no direct dependency between some models, such as the pavement settlement model and the waveform reflection and refraction model, they can be analyzed independently without affecting each other, and parallel execution can be selected. The series-parallel mixed execution is in some complex situations, where some models need to be executed in series, while the other part can be executed in parallel. For example, when analyzing crack extension, if the impact of wave propagation is small, the crack extension analysis can be performed first and executed in series, and then the settlement analysis and waveform reflection analysis can be performed in parallel. By clarifying the dependencies between the models, it is determined which models need to be executed serially and which can be executed in parallel, thereby optimizing the analysis process and improving the overall computing efficiency.

[0062] Step 306: Determine the physical model usage guidance information according to the analysis model serial execution information and the analysis model parallel execution information.

[0063] Specifically, when the guidance information of the analysis model serial execution information, the analysis model parallel execution information, and the series-parallel hybrid execution information are integrated into a unified physical model use guidance information, it is necessary to clearly define the dependencies and execution logic between each physical model. For the serial execution part, the guidance information will ensure that the output of the previous model is used as the input of the next model in order, and make clear guidance on the execution order; for the parallel execution part, the guidance information allows related models to be analyzed at the same time, and after the calculation is completed, the respective results are independently integrated into the final evaluation framework. For the series-parallel hybrid execution, it is necessary to flexibly adjust according to factors such as damage type, interaction between models, and computational efficiency. Some models may need to be executed in series first to obtain the necessary pre-results, while other independent models can be executed in parallel. Finally, these guidance information are integrated into a unified execution framework. The physical model use guidance information obtained not only ensures the execution order and data dependency integrity of each model, but also can flexibly allocate series, parallel and hybrid execution strategies.

[0064] Step 308: determining physical model setting parameters according to the physical model usage guidance information and the preliminary structural damage detection data.

[0065] Specifically, in the serial execution mode, since each physical analysis model is executed in a strict order, the output of each model will become the input of the next model. Therefore, in this mode, the preliminary structural damage detection data first provides the setting parameters for the first model, and the subsequent models will adjust their settings according to the results of the previous model. For example, the crack propagation model first sets the initial parameters of the crack according to the preliminary detection data (such as the initial position, morphology and depth of the crack), including the initial size, propagation rate, initial depth, etc. of the crack, which will affect the input of the subsequent elastic wave propagation model. In the elastic wave propagation model, parameters such as the propagation speed, path, and attenuation characteristics of the wave will depend on the characteristics of the crack, so the results of the crack propagation model will provide specific crack information for the wave propagation model. Through serial execution, the damage detection data is gradually passed to the next model, ensuring that each step of the analysis can inherit the damage information of the previous step, so as to obtain accurate damage assessment results.

[0066] In parallel execution mode, multiple physical analysis models run independently and do not affect each other. Therefore, in this mode, the preliminary structural damage detection data will provide independent setting parameters for each model. For example, the pavement settlement mechanical model and the waveform reflection and refraction model can be calculated simultaneously. The parameters of the former may include load distribution, initial thickness of the settlement area, support characteristics of the underground layer, etc., while the latter may include elastic modulus of the material, wave velocity, etc. Since there is no interdependence between the models, the setting parameters of each model are independently extracted from the preliminary structural damage detection data to ensure that the calculation of each model can be completed independently. This can improve the calculation efficiency and allow each model to analyze the pavement damage from different angles.

[0067] In the series-parallel hybrid execution mode, some physical analysis models need to be executed in series, while some physical analysis models can be executed in parallel. In this case, the parameter settings of the physical analysis models will be flexibly adjusted according to the interdependence between the models. For example, the crack extension model and the elastic wave propagation model need to be executed in series, and the output parameters of the crack extension model (such as the size and extension speed of the crack) will be used as the input of the elastic wave propagation model. In the parallel execution of the pavement settlement mechanics model and the waveform reflection and refraction model, the two models can run independently, so their setting parameters (such as the initial thickness of the settlement area, the underground support characteristics, and the elastic modulus of the pavement material, the wave propagation speed, etc.) are also independently obtained from the preliminary structural damage detection data without interfering with each other. Finally, all the model parameters of all these different execution modes are reasonably integrated according to the guidance information used by the physical model to ensure that the input of each model is accurate and consistent with the execution mode. Through this hybrid execution mode, it is possible to reasonably allocate computing resources while ensuring that each damage analysis process can take into account specific physical influences and relationships.

[0068] In this embodiment, by determining the appropriate physical analysis model based on the preliminary structural damage detection data and reasonably arranging its execution order, efficient collaborative analysis between different damage types can be achieved. This guidance information based on serial, parallel and hybrid execution strategies helps to ensure the accuracy and consistency of the input data of each physical model, thereby improving the accuracy and reliability of the overall analysis. At the same time, by optimizing the execution order and setting parameters of the analysis model, the actual damage situation can be simulated more accurately, providing a more scientific and accurate basis for pavement repair and maintenance decisions.

[0069] In an exemplary embodiment, Figure 4 As shown, according to the physical model use guidance information and physical model setting parameters, each physical analysis model is used to perform mechanical analysis and wave propagation analysis on the road surface to be tested, and the road surface structure damage detection data of the road surface to be tested is obtained, including steps 402 to 408. Among them:

[0070] Step 402, based on the parallel execution information of the analysis model, the elastic wave propagation model and the waveform reflection and refraction model are called to perform deep damage analysis to obtain road surface deep damage analysis data.

[0071] Among them, deep damage analysis can be performed by simulating the propagation characteristics of sound waves or radar waves in the road surface to analyze the damage in the deep area inside the road surface. The elastic wave propagation model and the waveform reflection and refraction model are used to analyze the impact of deep damage such as cracks and settlement on wave propagation, such as propagation speed, path, attenuation and other characteristics.

[0072] Among them, the pavement deep damage analysis data may be the specific results obtained through the deep damage analysis, including information such as the type, location, depth, and expansion trend of the pavement deep damage.

[0073] Specifically, when conducting deep damage analysis, the information is executed in parallel according to the analysis model, and the elastic wave propagation model and the waveform reflection and refraction model are started at the same time for analysis. The elastic wave propagation model simulates the propagation behavior of sound waves or radar waves in the road surface, and analyzes in detail the impact of deep damage such as cracks or settlement on wave propagation characteristics. Specifically, the elastic wave propagation model calculates indicators such as wave propagation speed, path, and energy attenuation. These indicators reflect the impact of deep structural damage on wave propagation and can reveal the depth and location of the damage. At the same time, the waveform reflection and refraction model analyzes the reflection and refraction of waves at the interfaces of different materials, thereby providing further information about the impact of the damaged area on wave propagation, including the reflection angle and refraction characteristics of the wave. By integrating the data analyzed by the elastic wave propagation model and the waveform reflection and refraction model, the deep damage analysis data of the road surface is obtained.

[0074] Among them, the expression of the elastic wave propagation model is:

[0075]

[0076] in, is the displacement vector field, To output dynamically corrected material stiffness tensor based on deep learning, For over time and location The effective density varies, is the local damage factor or material degradation coefficient, Excited by external force or wave source;

[0077] And, the expression of the waveform reflection and refraction model is,

[0078]

[0079]

[0080] in, Degradation of dependent materials The reflection coefficient, Degradation of dependent materials The transmission coefficient, is the incident angle, is the reflection coefficient under ideal undamaged or reference conditions, is the transmission coefficient under ideal undamaged or reference conditions, For over time and location Varying reflection correction function, For over time and location Varying transmission correction function, To calculate the results of layered propagation, For the Layer in time The wave pressure or displacement amplitude.

[0081] Step 404, based on the serial execution information of the analysis model, the crack propagation model and the pavement settlement mechanical model are called to perform serial morphological damage analysis to obtain serial pavement morphological damage analysis data.

[0082] Among them, serial morphological damage analysis can be to execute multiple analysis models in a specific order to simulate and evaluate the impact of pavement morphological damage (such as crack expansion, settlement, etc.) on the pavement structure. In the serial analysis process, the output results of the previous model will be used as the input of the subsequent model to ensure that the causal relationship between the models is accurately reflected.

[0083] The serial pavement morphological damage analysis data may be the result obtained through the serial morphological damage analysis, which usually includes the joint output of the crack propagation model and the pavement settlement mechanical model. The data reflects the influence of crack propagation on morphological damage such as pavement settlement and deformation.

[0084] Specifically, in the process of serial morphological damage analysis, the crack propagation model and the pavement settlement mechanical model are executed in a strict order. The crack propagation model will analyze the preliminary structural damage detection data (such as the location, morphology, depth, etc. of the cracks) to simulate the expansion of cracks under different loads. The crack propagation model will output the speed, direction and possible evolution path of crack propagation, which will be used as input for the next step of analysis. Next, the pavement settlement mechanical model analyzes the impact of crack propagation on pavement settlement based on the preliminary results of crack propagation. The model considers the deformation of the pavement structure after crack propagation, simulates the settlement changes under load, and analyzes factors such as the support characteristics of the underground layer. Since the crack propagation model and the settlement mechanical model are executed serially, the input of the latter is the output of the former, which ensures the logical continuity and interdependence between the two, making the assessment of pavement morphological damage more accurate.

[0085] Step 406, based on the parallel execution information of the analysis model, the crack propagation model and the pavement settlement mechanical model are called to perform parallel morphological damage analysis, to obtain parallel pavement morphological damage analysis data, and to obtain serial pavement morphological damage analysis data.

[0086] Parallel morphological damage analysis can be the simultaneous execution of multiple analysis models without mutual dependence, which is used to evaluate the pavement morphological damage (such as crack growth, settlement, etc.). During the parallel execution process, the crack growth model and the pavement settlement mechanical model are independently operated, each based on preliminary test data, and do not need to rely on the results of other models.

[0087] The parallel pavement morphological damage analysis data can be the result obtained through the parallel morphological damage analysis, including the crack propagation data and settlement data analyzed independently. The data is independent of other models, is not affected by serial execution, and reflects the different damage characteristics of crack and settlement areas.

[0088] Specifically, in the parallel morphological damage analysis, the crack propagation model and the pavement settlement mechanical model are independently executed simultaneously. The crack propagation model independently analyzes the morphology, size, propagation speed and possible evolution path of the cracks based on the preliminary detection data to obtain the crack propagation trend; at the same time, the pavement settlement mechanical model also runs independently to analyze the settlement characteristics of the pavement under load, especially the settlement changes in the crack area. Since the two models are executed in parallel, they are not interdependent and work independently, so they can quickly generate data on pavement morphological damage from multiple perspectives. The parallel execution mode can improve computational efficiency, avoid possible time delays in serial analysis, and comprehensively evaluate the different effects of crack propagation and settlement deformation on the pavement structure to obtain serial pavement morphological damage analysis data.

[0089] The expression of the crack extension model is:

[0090]

[0091] in, is the non-local crack damage field, For over time and location The effective stress intensity factor associated with the material degradation caused by the change, For over time and location Non-local operations on the interaction of changing cracks in a neighborhood, is the damage growth law function, is an external load or boundary condition;

[0092] And, when the crack propagation model and the pavement settlement mechanical model are executed in series, the expression of the pavement settlement mechanical model is:

[0093]

[0094] in, For over time and location The amount of sedimentation changes, For over time and location Changes in the dynamic elastic modulus associated with material degradation, Traffic load and environmental factors;

[0095] Alternatively, when the crack propagation model and the pavement settlement mechanical model are executed in series, the expression of the pavement settlement mechanical model is:

[0096]

[0097]

[0098] in, For over time and location The dynamic elastic modulus after the change causes material degradation and coupled crack damage field, is the damage influence coefficient, is the non-local crack damage field, Traffic load and environmental factors.

[0099] Step 408, performing decision analysis based on the pavement deep damage analysis data, the serial pavement morphology damage analysis data, and the parallel pavement morphology damage analysis data to obtain pavement structure damage detection data.

[0100] Among them, decision analysis can be the process of comprehensively processing all model analysis results to obtain the final pavement structure damage detection data. At this stage, the system will integrate deep damage analysis data, serial morphological damage analysis data, and parallel morphological damage analysis data to evaluate the type, location, depth, and expansion trend of different types of damage.

[0101] Specifically, after the analysis results of all physical models are completed, the system enters the decision analysis stage, at which time the deep damage analysis data, serial morphological damage analysis data and parallel morphological damage analysis data will be comprehensively considered. The system first identifies the type of damage (such as cracks, settlement, etc.) by fusing the damage information output by each model, and then determines the location, depth, size and possible development trend of the damage. Then the decision analysis is weighted by fusing the analysis results of different models, and considering the relationship between various types of damage to prioritize and obtain the pavement structure damage detection data.

[0102] In this embodiment, by reasonably utilizing parallel and serial execution strategies to schedule different physical analysis models, the advantages of each model can be fully utilized to achieve accurate analysis of deep damage and surface morphological damage to the road surface; parallel analysis can simultaneously process multiple damage features to improve analysis efficiency, while serial analysis ensures accurate modeling of complex damage interactions. Finally, by integrating various analysis results for decision analysis, detailed information on road structure damage can be obtained comprehensively and accurately, providing a scientific basis for subsequent repair and maintenance, and optimizing the accuracy and work efficiency of road damage detection.

[0103] In an exemplary embodiment, Figure 5 As shown, dual-channel structural damage identification is performed on the image detection data, radar detection data, and acoustic wave detection data to obtain preliminary structural damage detection data of the road surface to be detected, including steps 502 to 514. Among them:

[0104] Step 502: using the road surface damage identification channel, based on the image detection data and the radar detection data, identify the surface anomaly of the road surface to be detected, and obtain the road surface damage data.

[0105] Among them, the road surface damage identification channel can be a way to identify road surface damage mainly through image detection data and radar detection data. Image data captures surface cracks, potholes, wear and tear and other visible damage through visual perception, while radar data further helps identify shallow damage by measuring the intensity changes of road surface reflection waves.

[0106] Among them, the road surface damage data can be the result obtained through the road surface damage identification channel, including information on surface damage such as cracks, potholes, wear and tear, and pothole edges.

[0107] Specifically, in the road surface damage identification channel, image detection data and radar detection data are combined to identify and detect damage to the road surface. Image detection data provides visual information about road surface damage (such as cracks, potholes, peeling, etc.), while radar detection data reveals structural deformations such as settlement and crack depth by acquiring the morphological characteristics of the road surface. Through convolutional neural networks (CNN), these two types of data are combined to extract features of surface damage; CNN can effectively identify texture and shape features in images, and combined with the depth information provided by radar data, comprehensively analyze the morphology, location and size of damage, realize accurate identification of road surface damage, and obtain specific surface damage data, including the specific location, morphology and severity of the identified damage.

[0108] Step 504 , using the road structure damage identification channel, based on the radar detection data and the sound wave detection data, identifies the structural anomaly of the road surface to be detected, and obtains the road structure damage data.

[0109] Among them, the road structure damage identification channel can be a way to analyze the deep structural damage of the road surface through radar detection data and sound wave detection data. Radar data is used to detect the deformation of underground structures, such as settlement and crack depth; while sound wave data reveals the impact of deep damage on wave propagation through propagation characteristics, helping to identify deep damage to the road surface.

[0110] Among them, road structure damage data can be obtained through the road structure damage identification channel, which mainly includes deep damage information such as crack depth, settlement section, and wave propagation anomaly.

[0111] Specifically, in the road structure damage identification channel, radar detection data and acoustic wave detection data are mainly relied on to identify the deep structural damage of the road surface. Radar detection data provides detailed information about the deep deformation, settlement and crack extension of the road surface, while acoustic wave data reveals the propagation characteristics of waves in the road surface through time series changes, which can reflect the dynamic process of deep damage such as crack extension or settlement. Radar and acoustic wave data are fused through methods such as temporal convolutional networks (TCNs), making full use of the spatiotemporal information of both to identify the type, location and development trend of deep damage. Among them, radar provides spatial morphological data, and acoustic waves provide dynamic information of the damage development process, which can accurately identify structural damage, such as settlement zones, crack extension areas, etc., and obtain road structure damage data.

[0112] Step 506 , identifying the spatial environment information and data timing information of the road surface to be detected based on the image detection data, radar detection data, and sound wave detection data.

[0113] The spatial environment information may be spatial feature data related to the road surrounding environment, such as terrain, vegetation, buildings, traffic conditions, etc.

[0114] The data timing information may be time series data extracted from acoustic wave detection data, which reflects the dynamic evolution of road surface damage.

[0115] Specifically, the system identifies environmental features around the road by analyzing image detection data and radar detection data. Image data is used to extract visual information around the road, such as obstacles, buildings, vegetation, etc. around the road surface. Through image segmentation and feature extraction, the system can effectively capture the spatial structure of these environments, while radar data provides perception of underground structures through the intensity and delay of reflected signals, helping to analyze groundwater levels, geological layers or subsidence phenomena, etc. Combining these two types of data can fully understand the spatial environmental information around the road surface.

[0116] The system extracts dynamic changes related to road damage by analyzing the time series information in the sound wave detection data. The time characteristics of the sound wave signal during the propagation process, such as the propagation speed, attenuation characteristics and changes in the reflected wave, reflect the real-time changes of the road surface or its surrounding environment. For example, if the sound wave propagation characteristics change significantly during a certain period of time, it may indicate new damage to the road surface or the expansion of existing damage, and obtain data time series information.

[0117] By combining the timing information of acoustic wave data with image and radar data, the system can infer the occurrence, expansion and dynamic change trend of damage in the time dimension, providing important spatiotemporal background for further damage analysis.

[0118] Step 508: Based on the spatial environment information, spatial feature fusion is performed on the road surface damage data and the road structure damage data to obtain spatial fusion detection data.

[0119] Among them, spatial feature fusion can be the process of combining road surface damage data and structural damage data with spatial environmental information to explore the spatial distribution characteristics of road surface damage.

[0120] Among them, the spatial fusion detection data can be the output data obtained by combining road surface damage data, road structure damage data and spatial environment information.

[0121] Specifically, in the process of spatial feature fusion, geographic information system (GIS) technology is used to align the spatial positions of road surface damage data and road structure damage data to ensure that all damage data can be correctly mapped to the same spatial coordinate system. Furthermore, the key step of spatial feature fusion is to spatially combine surface damage data with deep structure damage data, which mainly relies on spatial mapping technology, which takes into account the spatial characteristics of different damage types, such as the relative position and morphology of surface cracks and the extension range of deep cracks. By integrating these data, spatial fusion can intuitively show which surface damage and deep damage have intersections or mutual influences, thereby providing a more comprehensive perspective for subsequent damage analysis and obtaining spatial fusion detection data.

[0122] Step 510, based on the data time series information, time series feature fusion is performed on the road surface damage data and the road structure damage data to obtain time series fusion detection data.

[0123] Among them, time series feature fusion can be a process of combining road surface damage data and road structure damage data with data time series information.

[0124] The time series fusion detection data may be a data set obtained by fusing road surface damage data, road structure damage data and time series information.

[0125] Specifically, in the process of time series feature fusion, time series features are first extracted from each damage data source. For example, road surface damage data reflects the size and position changes of cracks on the road surface at different time points, while road structure damage data may reveal the expansion speed of deep cracks, the change speed of settlement areas, etc. The time series of each data source will be decomposed into damage indicators at multiple time points, such as the change trend of crack width, depth, morphology, etc. over time.

[0126] Next, the system will analyze the rate of change of surface damage and deep structural damage in each time period, identifying which surface damage is accelerated under the influence of structural damage, and which deep damage may cause the expansion of surface damage in the future. Time series feature fusion can not only reveal the dynamic relationship between different types of damage, but also provide a reference for future damage prediction. For example, if the expansion rate of surface cracks in a certain section of pavement is highly correlated with the expansion rate of deep cracks, the system can infer that the expansion of deep cracks is the root cause of the aggravation of surface cracks, and vice versa, and finally obtain time series fusion detection data.

[0127] Step 512, determining the spatiotemporal dependency of the detection data from the spatial fusion detection data and the temporal fusion detection data.

[0128] The spatiotemporal dependency of detection data may refer to the correlation and interdependence between damage data in spatial and temporal dimensions.

[0129] Specifically, since the spatiotemporal dependency of the data can reveal the mutual influence and relationship between different damage data in space and time, based on the spatial fusion detection data and the temporal fusion detection data in the previous steps, the potential mutual influence between the two data is further analyzed by spatiotemporal correlation. That is, by calculating the correlation between the spatial coordinates and time series of the spatial fusion detection data and the temporal fusion detection data, the possible causal relationship between surface damage and structural damage is identified, and the spatiotemporal dependency of the detection data is determined.

[0130] Specifically, the system will analyze the spatial adjacent relationship of the spatial fusion data to identify which damage locations have common spatial characteristics or correlations (for example, whether the expansion of cracks is accompanied by changes in the settlement area); it will also analyze the temporal trend of the time series fusion detection data to find out which surface damage expansion is synchronous or delayed with the change trend of structural damage. Through this spatiotemporal correlation analysis, the system can capture the coupling effects between different damage types in time and space, identify which damage expansion in a specific period of time may lead to other damage, or which areas of damage will show a trend of acceleration or further deterioration at some point in the future, and determine the spatiotemporal dependency of the detection data.

[0131] Step 514, based on the spatiotemporal dependency of the detection data, the spatial fusion detection data and the temporal fusion detection data are fused using a spatiotemporal attention mechanism to obtain preliminary structural damage detection data.

[0132] Among them, the spatiotemporal attention mechanism fusion can be a process of weighted fusion of spatial fusion detection data and temporal fusion detection data based on spatiotemporal dependency.

[0133] Specifically, based on the identified spatiotemporal dependency of the detection data, weights are assigned to the damage in different regions and time periods in the spatial fusion detection data and the temporal fusion detection data. For example, cracks in certain locations may have a greater impact on the overall structure of the pavement, or damage changes in certain time periods have stronger predictive significance. This information is given a higher weight through the spatiotemporal attention mechanism. Specifically, the spatiotemporal attention mechanism assigns different weights to different spatiotemporal regions by calculating the importance distribution of damage data in space and time. It can autonomously adjust the model's attention to different types of damage and give priority to those damaged areas that have a greater impact on the overall pavement structure. The spatiotemporal attention mechanism not only enhances the effectiveness of damage data, but also improves the ability to predict future damage evolution trends. Finally, after fusion through the spatiotemporal attention mechanism, the system obtains preliminary structural damage detection data, which contains key information such as the type, location, and development trend of various types of pavement damage.

[0134] In this embodiment, by combining image, radar and sound wave detection data and adopting a multi-channel recognition and fusion method, the road surface and structural damage can be accurately identified and analyzed while taking into account the spatial environment and temporal changes; the spatial feature fusion and temporal feature fusion technology further improves the spatial accuracy and temporal consistency of damage detection, and optimizes the correlation between different damage types through the introduction of spatiotemporal dependency analysis and spatiotemporal attention mechanism, thereby effectively improving the accuracy and comprehensiveness of preliminary structural damage detection, and providing reliable data support for subsequent road maintenance and reinforcement.

[0135] In an exemplary embodiment, Figure 6 As shown, after the steps of using the guidance information of the physical model and setting the parameters of the physical model, respectively performing mechanical analysis and wave propagation analysis on the road surface to be tested and obtaining the road surface structure damage detection data of the road surface to be tested, the method further includes steps 602 to 616. Among them:

[0136] Step 602, predicting the future usage intensity of the road surface to be detected based on the historical road surface traffic flow data and future weather forecast data of the road surface to be detected, and obtaining road surface traffic flow forecast data.

[0137] The historical traffic data of the road surface may include the number of vehicles, vehicle types, vehicle speeds, traffic density, etc., which pass on the road surface within a period of time. These data may be obtained through traffic monitoring systems, vehicle sensors, or road sensors.

[0138] Among them, future weather forecast data can be the weather conditions in a certain period of time in the future predicted by using meteorological models and historical meteorological data, combined with satellite monitoring, ground meteorological stations and other meteorological monitoring methods, including temperature, precipitation, wind speed, humidity and other information.

[0139] Among them, future usage intensity can be a comprehensive analysis of the road's historical traffic data and future weather forecast data to predict the traffic load intensity that the road will bear in a certain period of time in the future, reflecting the comprehensive impact of traffic volume, vehicle speed, traffic type and weather factors that the road will face in a certain period of time in the future.

[0140] Among them, the road traffic flow prediction data can be a combined analysis of the road's historical traffic flow data and future weather forecast data to predict the changing trends of future road traffic flow, vehicle speed, and traffic patterns.

[0141] Specifically, first obtain the historical traffic data of the road surface to be tested, which includes the daily traffic flow, traffic peak hours, and the impact of special events (such as holidays, road repairs, etc.) on the traffic flow of the road surface to be tested in the past period of time. Then combine the future weather forecast data, which includes expected precipitation, temperature changes, wind force and other factors, and use the prediction model for analysis; since weather conditions usually have a significant impact on traffic flow, for example, heavy rain may cause a decrease in vehicle speed and traffic flow, or snowy days may cause road closures and traffic interruptions. After comprehensively considering these variables, the historical road traffic data is combined with future weather forecast data through time series analysis methods (such as time series analysis, regression models, machine learning models, etc.) to predict the trend of traffic changes in the future period of time. The model will obtain the road traffic forecast data based on the seasonal changes and periodic fluctuations identified in the historical data, combined with the prediction of future conditions in the weather forecast. The road traffic forecast data includes the expected traffic peak period, trough period and future road load conditions.

[0142] Step 604, determining the abnormal road surface repair process for the road surface to be inspected based on the road traffic prediction data, future weather prediction data and road structure damage detection data.

[0143] Among them, the abnormal road surface repair process can be based on the road surface damage type (such as cracks, potholes, settlement, etc.) and the severity of the damage, combined with the road surface historical traffic flow, weather forecast data, etc., to select appropriate repair methods and technical means. For example, for severe cracks, the crack injection repair process may be used, while for settlement problems, the repair method of strengthening the road surface foundation may be selected.

[0144] Specifically, since traffic forecast data can reflect the intensity of road use in the future, it helps to determine whether the road surface will be overloaded with traffic pressure. For example, some roads are expected to bear higher traffic volume, which may lead to more surface damage or structural damage, thus requiring the selection of more durable and load-bearing repair processes. Future weather forecast data (such as expected rainfall, temperature changes, snowy days, etc.) will play a vital role in the selection of repair processes. For example, in cold or rainy areas, some repair processes (such as hot mix asphalt repair) may be affected by temperature and humidity, resulting in unsatisfactory repair results. Therefore, it is necessary to select materials and processes that adapt to these climatic conditions. In addition, pavement structure damage detection data provides the specific type and severity of damage, which provides a direct basis for the selection of repair processes. If the detection data shows that the structural damage is more serious, more complex and in-depth repair processes may be required, such as overall reinforcement or local overlay; while if it is just surface cracks or potholes, simple repairs may be sufficient.

[0145] In the specific implementation, the traffic density and peak traffic flow in the future, especially the proportion of heavy vehicles, are calculated through traffic forecast data to identify road sections with greater load-bearing pressure; for these high-flow areas, by evaluating the load-bearing capacity of existing pavement materials, more durable and high-load-bearing repair materials (such as high-strength modified asphalt or cement concrete) are selected. Combined with weather forecast data, calculate the impact of weather factors on construction, such as low temperature or high humidity may affect the curing speed or construction quality of repair materials; according to climatic conditions, choose repair processes that are suitable for severe weather, such as fast-curing materials or special materials for low-temperature construction. Based on the pavement structure damage detection data, analyze the type and severity of the damage, such as the depth and width of the cracks, or whether there is structural settlement, and determine the repair plan based on the specific characteristics of the damage; for sections with more serious structural damage, deep repair or reinforcement processes may be required, while for surface cracks or mild damage, a simpler repair process is selected. By combining the above different data and inputting them into the process optimization algorithm, further process optimization calculations are performed on different sections of the road to be tested, and the repair needs and required repair process data of different sections are comprehensively obtained, and then the road abnormality repair process of the road to be tested is determined according to each repair process data.

[0146] Step 606, determining the road surface abnormality repair material from the preset repair materials for the road surface to be inspected according to the road surface abnormality repair process and future weather forecast data.

[0147] Among them, road surface abnormality repair materials can be special materials selected for repairing different types of road damage (such as cracks, potholes, settlement, etc.). According to different types of damage, future traffic flow, weather forecast and other data, the system will select appropriate repair materials, such as high-strength concrete, repair glue, plastic filler, etc. The material selection criteria include durability, compression resistance, water resistance and other performance requirements to ensure the long-term stability of the repair effect.

[0148] Specifically, materials need to be screened according to the requirements of the abnormal road surface repair process. For example, if the repair process requires structural reinforcement or deep repair, the material selection will focus on materials with strong bearing capacity and durability, such as cement-based materials or high-strength modified asphalt, to ensure that the repaired road surface can withstand high traffic volume or long-term use by heavy vehicles. If it is for the repair of surface cracks or small area damage, more economical and convenient construction materials such as cold mix asphalt or fast curing materials can be selected. Then, combined with future weather forecast data, analyze the climatic conditions during the repair period, adjust the above-mentioned screening of materials according to the requirements of the abnormal road surface repair process, and ensure that the selected materials can adapt to the expected weather. For example, if continuous rainfall or high humidity is expected, select materials with good waterproofness and curing performance in a humid environment; if low temperature weather is expected during construction, select repair materials that can quickly cure at low temperatures or whose curing performance is not affected by low temperatures. Since the adjustment of materials based on future weather forecast data does not take into account sudden weather changes, the adjustment of materials based on future weather forecast data is combined with the redundant values ​​of the impact of sudden weather changes, and the most suitable materials are selected from the preset repair materials as abnormal road surface repair materials to ensure that the repair materials can effectively exert their performance under different climatic conditions and ensure the durability and stability of the repair effect.

[0149] Step 608 , three-dimensional modeling is performed based on the road surface abnormality repair process and road surface abnormality repair materials, and initial construction area parameters of the road surface to be inspected are calculated.

[0150] The initial construction area parameters may be the preliminary construction area determined by the system according to the road damage data, construction requirements and preset repair plan during the road repair process. These parameters include the specific location, size, depth and required repair materials of the construction area.

[0151] Specifically, the requirements of the road surface repair process and the characteristics of the selected road surface repair materials are converted into specific construction parameters; for example, the road surface repair process requires deep reinforcement of severely damaged roads, or local repair of the surface, which will affect the depth and width of the construction area and the amount of repair materials required. In the 3D modeling process, the geographical location, morphology and damaged area of ​​the road to be inspected are accurately modeled through digital terrain scanning or laser radar technology to generate a detailed 3D model. Based on this model, the basic parameters such as the size, area, and volume of the repair area can be calculated, and the boundary and geometry of the repair area can be determined. In addition, according to the characteristics of the selected road surface repair materials (such as high-strength asphalt, cold mix asphalt or cement-based materials), combined with the construction process requirements, the thickness, number of layers and material proportions of each layer of the required materials are calculated, and the preliminary construction area parameters are obtained by combining the above data.

[0152] Step 610, input the road traffic flow prediction data into the route planning algorithm of the road to be detected to obtain initial road diversion data.

[0153] Among them, the route planning algorithm can calculate the best road diversion plan based on road traffic data, traffic patterns, construction areas, road diversions and other information.

[0154] The initial road diversion data may be a preliminary traffic diversion plan generated based on the planning of the construction area and the road traffic volume forecast data. The data includes the traffic flow route, detour path and diversion guidance during the construction period.

[0155] Specifically, the future traffic flow information obtained from the traffic flow prediction data, especially the traffic density and the proportion of heavy vehicles during peak hours, is used to evaluate the traffic pressure of the road section to be repaired during the construction period. Based on this information, the route planning algorithm will analyze which sections of the road may cause traffic bottlenecks or congestion due to the repair work, and then automatically generate a set of alternative routes. Among them, the route planning algorithm usually combines factors such as real-time traffic flow, road carrying capacity, intersection conditions, road grade, and the specific location of the repair area to optimize the road diversion plan. Specifically, the route planning algorithm determines the optimal traffic diversion path during the construction period by simulating different traffic flow distributions and diversion plans to minimize traffic interference and improve construction efficiency. Finally, through the calculation of the route planning algorithm, the initial road diversion data is generated, including specific information such as the diverted route, intersection adjustment, and signal light configuration, and the dynamic changes of traffic flow during the construction period are taken into account.

[0156] Step 612, optimizing the initial construction area parameters and the initial road diversion data to obtain the target construction area data and the target road diversion data.

[0157] Among them, the target road diversion data can be a diversion plan that has been further optimized and adjusted based on the initial road diversion data. These data are processed by an optimization algorithm, taking into account factors such as the impact of construction on traffic, changes in traffic flow, and construction safety.

[0158] Specifically, first evaluate whether the preliminary construction area and diversion plan can effectively meet the construction needs and minimize traffic impact. At this stage, the optimization algorithm will consider multiple factors for optimization. For the construction area parameters, the preliminary determined construction area needs to be adjusted according to the actual situation of the road damage (such as depth, range) and the requirements of the repair process (such as deep reinforcement or surface repair) to ensure that the materials and processes used in the repair process can cover the damaged area to the greatest extent while avoiding excessive interference with the surrounding environment and traffic flow; this may include adjusting the size, shape or construction period of the repair area to complete the repair work more efficiently. For road diversion data, the optimization process will consider the traffic flow during construction, the road's capacity, available alternative routes, etc., to find the best diversion plan that can balance traffic flow, avoid congestion and maximize construction efficiency. This process is usually implemented through multi-objective optimization algorithms (such as genetic algorithms, simulated annealing, etc.) to ensure that the final plan not only meets the repair requirements, but also minimizes the impact on surrounding traffic. The final target construction area data and target road diversion data will include information such as optimized construction boundaries, diversion routes, construction periods, and traffic flow adjustments.

[0159] Step 614, based on the target construction area data and the target road diversion data, perform construction risk prediction on the road surface to be inspected to obtain construction risk prediction data.

[0160] Among them, construction risk prediction can be based on factors such as the geographical environment, traffic flow, weather forecast, construction technology, etc. of the construction area to predict various risks that may be encountered during the construction process. These risks include traffic accidents during construction, equipment failures, material supply shortages, bad weather, etc. Through the analysis of historical data and existing data, the system can identify potential construction risks in advance.

[0161] Among them, the construction risk prediction data can be risk data generated by quantitatively analyzing various risks that may occur during the construction process. These data include the probability of occurrence of different types of risks, the severity of the risks, and the scope of impact.

[0162] Specifically, the risk factors that may be faced during the construction process are evaluated by combining the geographical characteristics of the target construction area data, construction depth, repair technology and traffic flow changes after the target road diversion data. Construction risk prediction usually includes the following aspects: traffic congestion, construction safety, environmental impact and material performance. The specific process predicts the risk of traffic congestion or accidents that may be caused by construction by analyzing the changes in traffic flow in the target construction area, especially the traffic density on the diversion route, traffic signal changes and the traffic capacity of the intersection; further evaluate the possible construction safety risks during the construction process, such as road stability, the use of mechanical equipment, and the clarity of safety signs; then analyze the impact of environmental factors (such as weather, temperature, humidity, etc.) on the construction progress and the performance of repair materials during the construction period, especially under extreme weather conditions, the construction risk may increase, such as low temperature may cause incomplete curing of materials or delays in the construction progress; finally, according to the road diversion plan and the specific conditions of the construction area, predict the potential risks of construction delays and cost overruns. By integrating these risk data and risk information, prediction models (such as risk assessment models, Monte Carlo simulations, etc.) are used to quantitatively analyze the risks that may occur during the construction process, and finally obtain construction risk prediction data.

[0163] Step 616, the road surface abnormality repair process, road surface abnormality repair materials, target construction area data, target road diversion data and construction risk prediction data are integrated to obtain road maintenance guidance information for the road surface to be inspected.

[0164] Among them, the road maintenance guidance information can be a comprehensive guidance document generated by comprehensively considering information on repair technology, repair materials, construction area, road diversion plan, construction risk prediction, etc. This information helps the construction party to scientifically and reasonably plan resources, arrange construction sequence, and manage risks during the road repair process.

[0165] Specifically, the abnormal road surface repair process, abnormal road surface repair materials, target construction area data, target road diversion data, and construction risk prediction data are integrated to form a comprehensive construction guidance framework. Specifically, the data of abnormal road surface repair process and abnormal road surface repair materials provide the technical requirements and material selection schemes for repair, the target construction area data and target road diversion data provide the spatial scope of construction and traffic diversion strategies, and the construction risk prediction data help identify potential construction challenges and safety hazards. In the fusion process, each data source is uniformly formatted through the data integration platform for subsequent analysis and processing. Then, through multi-dimensional analysis methods (such as data fusion algorithms, decision support systems, etc.), the selection of repair processes and materials is combined with the specific conditions of the construction area to ensure that the repair process can be effectively implemented in the predetermined area and minimize the interference with traffic flow. At the same time, the construction risk prediction data will be used to quantify the possible safety hazards, delay risks, cost control, etc. during the construction process, thereby affecting the adjustment of the construction plan and ensuring the feasibility and safety of the construction plan. Finally, combined with this information, detailed road maintenance guidance information is generated, including the type of repair materials, construction steps, construction period, diversion route, traffic diversion measures, risk warning and emergency response plans.

[0166] In this embodiment, by comprehensively analyzing the historical traffic data of the road surface, future weather forecast data, and structural damage detection data, this method can accurately predict the future usage intensity and traffic volume of the road surface, thereby providing a scientific basis for road surface repair. According to the traffic forecast data and weather information, combined with the structural damage detection results, the repair process selection can be optimized, and on this basis, the appropriate repair materials can be determined. Through three-dimensional modeling, the parameters of the initial construction area are accurately calculated, and the route planning algorithm is combined to generate diversion data to ensure smooth traffic during construction. After optimizing the construction area and diversion plan, construction risk prediction is further performed to ensure safety and efficiency during the construction process. Finally, by integrating the road maintenance guidance information generated by the above data, comprehensive support is provided for repair decisions to ensure the efficiency and accuracy of maintenance work.

[0167] Based on the same inventive concept, the present application also provides a pavement structure damage detection device for implementing the pavement structure damage detection method mentioned above. Figure 7 As shown, it includes: a road surface detection data acquisition module 702, an artificial intelligence analysis module 704, a physical parameter determination module 706 and a physical model analysis module 708.

[0168] In an exemplary embodiment, an electronic device is provided. The electronic device may be a server, and its internal structure diagram may be as shown in FIG. Figure 8 As shown. Those skilled in the art can understand that Figure 8The structure shown in the figure is merely a block diagram of a partial structure related to the scheme of the present application, and does not constitute a limitation on the electronic device to which the scheme of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0169] In one embodiment, a computer program product or computer program is also provided, the computer program product or computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs the steps in the above-mentioned method embodiments.

[0170] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0171] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. For ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of this application, which all belong to the protection scope of this application. Therefore, the protection scope of this application shall be based on the attached claims.

Claims

1. A method for detecting pavement structure damage, characterized in that: The method comprises: Acquire image detection data, radar detection data, and sound wave detection data of the road surface to be detected; Performing dual-channel structural damage identification on the image detection data, the radar detection data, and the sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected; Determining, based on the preliminary structural damage detection data, the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected; Wherein, determining the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected according to the preliminary structural damage detection data includes: Determining at least two physical analysis models corresponding to the road surface to be inspected according to the preliminary structural damage detection data; Determining analysis model serial execution information and analysis model parallel execution information according to the correlation between each of the physical analysis models; Determining the physical model usage guidance information according to the analysis model serial execution information and the analysis model parallel execution information; Determining the physical model setting parameters according to the physical model usage guidance information and the preliminary structural damage detection data; According to the guidance information for using the physical model and the parameters set for the physical model, each physical analysis model is used to perform mechanical analysis and wave propagation analysis on the road surface to be detected, so as to obtain the pavement structure damage detection data of the road surface to be detected.

2. The method according to claim 1, characterized in that The physical analysis model includes an elastic wave propagation model, a waveform reflection and refraction model, a crack extension model, and a pavement settlement mechanical model; the method of using the guidance information and the physical model setting parameters according to the physical model, using each physical analysis model to perform mechanical analysis and wave propagation analysis on the road surface to be detected, and obtaining the pavement structure damage detection data of the road surface to be detected, includes: According to the parallel execution information of the analysis model, the elastic wave propagation model and the waveform reflection and refraction model are called to perform deep damage analysis to obtain pavement deep damage analysis data; According to the serial execution information of the analysis model, calling the crack propagation model and the pavement settlement mechanical model to perform serial morphological damage analysis to obtain serial pavement morphological damage analysis data; And, according to the parallel execution information of the analysis model, calling the crack propagation model and the pavement settlement mechanical model to perform parallel morphological damage analysis to obtain parallel pavement morphological damage analysis data; A decision analysis is performed based on the pavement deep damage analysis data, the serial pavement morphology damage analysis data, and the parallel pavement morphology damage analysis data to obtain the pavement structure damage detection data.

3. The method according to claim 2, characterized in that The expression of the elastic wave propagation model is: in, is the displacement vector field, To output dynamically corrected material stiffness tensor based on deep learning, For over time and location The effective density varies, is the local damage factor or material degradation coefficient, Excited by external force or wave source; And, the expression of the waveform reflection and refraction model is, in, Degradation of dependent materials The reflection coefficient, Degradation of dependent materials The transmission coefficient, is the incident angle, is the reflection coefficient under ideal undamaged or reference conditions, is the transmission coefficient under ideal undamaged or reference conditions, For over time and location Varying reflection correction function, For over time and location Varying transmission correction function, To calculate the results of layered propagation, For the Layer in time The wave pressure or displacement amplitude.

4. The method according to claim 2, characterized in that: The expression of the crack extension model is: in, is the non-local crack damage field, For over time and location The effective stress intensity factor associated with the material degradation caused by the change, For over time and location Non-local operations on the interaction of changing cracks in a neighborhood, is the damage growth law function, is an external load or boundary condition; And, when the crack propagation model and the pavement settlement mechanical model are executed in series, the expression of the pavement settlement mechanical model is: in, For over time and location The amount of sedimentation changes, For over time and location Changes in the dynamic elastic modulus associated with material degradation, Traffic load and environmental factors; Alternatively, when the crack propagation model and the pavement settlement mechanical model are executed in series, the expression of the pavement settlement mechanical model is: in, For over time and location The dynamic elastic modulus after the change causes material degradation and coupled crack damage field, is the damage influence coefficient, is the non-local crack damage field, Traffic load and environmental factors.

5. The method according to claim 1, characterized in that The performing dual-channel structural damage identification on the image detection data, the radar detection data, and the sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected includes: Using a road surface damage identification channel, identifying surface abnormalities of the road surface to be detected according to the image detection data and the radar detection data, and obtaining road surface damage data; Using a road structure damage identification channel, according to the radar detection data and the sound wave detection data, identifying the structural abnormality of the road surface to be detected, and obtaining road structure damage data; Identifying the spatial environment information and data timing information of the road surface to be detected according to the image detection data, the radar detection data and the sound wave detection data; According to the spatial environment information, spatial feature fusion is performed on the road surface damage data and the road structure damage data to obtain spatial fusion detection data; According to the data time series information, the road surface damage data and the road structure damage data are subjected to time series feature fusion to obtain time series fusion detection data; Determining the spatiotemporal dependency of the detection data from the spatial fusion detection data and the temporal fusion detection data; According to the spatiotemporal dependency of the detection data, the spatial fusion detection data and the temporal fusion detection data are fused using a spatiotemporal attention mechanism to obtain the preliminary structural damage detection data.

6. The method according to claim 1, characterized in that After the step of using the guidance information of the physical model and setting parameters of the physical model, calling the physical analysis model to perform mechanical analysis and wave propagation analysis on the road surface to be detected respectively, and obtaining the pavement structure damage detection data of the road surface to be detected, the method further includes: Predicting the future use intensity of the road surface to be detected based on the road surface historical traffic flow data and future weather forecast data of the road surface to be detected, and obtaining road surface traffic flow forecast data; Determining a road surface abnormality repair process for the road surface to be detected according to the road surface traffic flow prediction data, the future weather prediction data, and the road surface structure damage detection data; Determining a road surface abnormality repair material from preset repair materials for the road surface to be inspected according to the road surface abnormality repair process and the future weather forecast data; Performing three-dimensional modeling according to the road surface abnormality repair process and the road surface abnormality repair material, and calculating the initial construction area parameters of the road surface to be inspected; Inputting the road traffic flow prediction data into the route planning algorithm of the road to be detected to obtain initial road diversion data; Optimizing the initial construction area parameters and the initial road diversion data to obtain target construction area data and target road diversion data; According to the target construction area data and the target road diversion data, a construction risk prediction is performed on the road surface to be inspected to obtain construction risk prediction data; The road abnormality repair process, the road abnormality repair material, the target construction area data, the target road diversion data and the construction risk prediction data are integrated to obtain road maintenance guidance information for the road to be inspected.

7. A road structure damage detection device, characterized in that: The device comprises: A road surface detection data acquisition module is used to acquire image detection data, radar detection data and sound wave detection data of the road surface to be detected; An artificial intelligence analysis module, used for performing dual-channel structural damage identification on the image detection data, the radar detection data and the sound wave detection data to obtain preliminary structural damage detection data of the road surface to be detected; A physical parameter determination module, used to determine the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected according to the preliminary structural damage detection data; Wherein, determining the physical model usage guidance information and physical model setting parameters corresponding to the road surface to be detected according to the preliminary structural damage detection data includes: Determining at least two physical analysis models corresponding to the road surface to be inspected according to the preliminary structural damage detection data; Determining analysis model serial execution information and analysis model parallel execution information according to the correlation between each of the physical analysis models; Determining the physical model usage guidance information according to the analysis model serial execution information and the analysis model parallel execution information; Determining the physical model setting parameters according to the physical model usage guidance information and the preliminary structural damage detection data; The physical model analysis module is used to use the guidance information of the physical model and the parameters of the physical model, and use each physical analysis model to perform mechanical analysis and wave propagation analysis on the road surface to be detected, so as to obtain the road surface structure damage detection data of the road surface to be detected.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

Images