Method for evaluating residual performance of traffic protection facility after accident based on digital twinning

Abstract

The invention provides a traffic protection facility post-accident residual performance evaluation method based on digital twinning, and relates to the field of electric digital data processing, and the method comprises the steps: obtaining accident video data and vehicle parameter information uploaded by an accident vehicle; extracting collision characteristic parameters of the collision process in the accident video data; establishing a digital twin model under the accident condition according to the vehicle parameter information and the collision characteristic parameters, simulating a collision occurrence process, and obtaining stress distribution data and plastic deformation data of the protection facility; combining performance requirements in protection facility design specifications to establish a residual performance evaluation index system; and calculating the residual performance index of the traffic protection facility according to the residual performance evaluation index system, and matching a corresponding maintenance and reinforcement suggestion. By implementing the method, subjectivity and time delay of traditional manual detection are avoided, the residual performance state of the protection facility can be rapidly and accurately evaluated after an accident occurs, and evaluation efficiency and accuracy are improved.

Description

Technical Field

[0001] This application relates to the field of electronic digital data processing, and in particular to a method for evaluating the remaining performance of traffic protection facilities after an accident based on digital twins. Background Art

[0002] With the rapid development of China's transportation industry, the construction scale of important traffic infrastructure such as expressways and urban expressways has been continuously expanding. As an important facility to ensure road traffic safety, the integrity and reliability of traffic protection facilities are directly related to the life and property safety of road users. During the operation of the road, traffic protection facilities will inevitably encounter accidents such as vehicle collisions, which bring great challenges to the maintenance work of road operation units.

[0003] In related technologies, the performance evaluation of traffic protection facilities after being hit by vehicles mainly relies on manual on-site inspections. Inspectors visually check the apparent damages such as the deformation degree and cracking condition of the protection facilities, and make judgments on whether they need to be repaired or replaced based on experience. At the same time, some road operation units will also use portable detection instruments to quantitatively measure individual parameters such as the geometric dimension deviation and steel strength of the protection facilities as an auxiliary basis for evaluation.

[0004] However, the detection process of traditional evaluation methods is time-consuming. Inspectors need to observe and measure the damaged protection facilities section by section. At the same time, since the detection data mainly focuses on the visually apparent damaged parts, it is easy to miss the changes in the stress distribution inside the protection facilities and potential structural damages. Especially in accident-prone sections, frequent on-site detection operations not only increase the road traffic pressure but also easily cause the risk of secondary accidents. Thus, in the case of increasingly busy road traffic and continuously improving requirements for rapid response after accidents, the efficiency and accuracy of this traditional evaluation method are difficult to meet the actual needs. Summary of the Invention

[0005] This application provides a method for evaluating the remaining performance of traffic protection facilities after an accident based on digital twins, which is used to address the problem of how to quickly and accurately evaluate the remaining performance state of damaged traffic protection facilities after a collision accident.

[0006] In a first aspect, this application provides a method for evaluating the remaining performance of traffic protection facilities after an accident based on digital twins, which is applied to a traffic protection facility performance evaluation system. The method includes: Obtain accident video data and vehicle parameter information uploaded by the accident vehicle. The accident video data includes the complete process videos before, during, and after the accident, and the vehicle parameter information includes vehicle mass, vehicle size, and vehicle stiffness; Extract the collision characteristic parameters of the collision process in the accident video data through computer vision algorithms, where the collision characteristic parameters include collision speed, collision angle, collision position, and the deformation amount of the protective facilities; Establish a digital twin model under accident conditions based on the vehicle parameter information and the collision characteristic parameters. The digital twin model includes a vehicle kinematic model and a protective facility finite element model; Simulate the collision occurrence process in the digital twin model to obtain the stress distribution data and plastic deformation data of the protective facilities; Based on the stress distribution data and plastic deformation data, and combined with the performance requirements in the protective facility design specifications, establish a remaining performance evaluation index system. The remaining performance evaluation index system includes structural strength, geometric dimensions, and functional integrity; According to the remaining performance evaluation index system, calculate the remaining performance index of the traffic protective facilities, and match the corresponding maintenance and reinforcement suggestions based on the remaining performance index.

[0007] Through the above embodiments, the system can quickly obtain the stress distribution and deformation data of the protective facilities without on-site manual detection by acquiring accident video data and vehicle parameters, combining computer vision to extract collision characteristics, and establishing a digital twin model for simulation analysis. Based on these data, a remaining performance evaluation index system is established, which can comprehensively evaluate the structural strength, geometric dimensions, and functional integrity of the protective facilities and give corresponding maintenance suggestions. This method avoids the subjectivity and time delay of traditional manual detection, can quickly and accurately evaluate the remaining performance state of the protective facilities after an accident, and improves the evaluation efficiency and accuracy.

[0008] In some embodiments, the step of extracting the collision characteristic parameters of the collision process in the accident video data through computer vision algorithms specifically includes: Extract the video frame sequence from the accident video data and perform image enhancement and target segmentation processing to obtain the protective facility area image; Perform feature point detection and tracking on the protective facility area image to obtain feature point motion data; Based on the feature point motion data, use kinematic analysis methods to calculate the collision speed and collision angle, and use three-dimensional reconstruction methods to determine the collision position and the deformation amount of the protective facilities.

[0009] Through the above embodiments, the system can accurately obtain key parameters such as speed, angle, position, and deformation amount during the collision process through computer vision technologies such as video frame sequence processing, feature point detection and tracking, and 3D reconstruction. This method replaces the traditional way that requires manual measurement and estimation, not only improving the efficiency of data acquisition but also enhancing the accuracy of parameter measurement. These high-precision collision feature parameters provide reliable input data for the subsequent digital twin model.

[0010] In some embodiments, after the step of detecting and tracking feature points in the image of the protection facility area to obtain feature point motion data, the following steps are further included: Extract audio data from the accident video data and perform noise reduction and segmentation processing to obtain collision acoustic signals; Perform time-frequency analysis on the collision acoustic signals to obtain a spectrogram, and then extract acoustic feature parameters from the spectrogram. The acoustic feature parameters include the peak frequency and energy distribution of the collision sound wave; Filter and optimize the feature point motion data based on the acoustic feature parameters to obtain optimized feature point motion data.

[0011] Through the above embodiments, the system optimizes the visual feature point motion data by introducing collision acoustic signal analysis, effectively improving the accuracy of collision feature parameters. Acoustic features can capture collision details that are difficult to obtain visually, such as internal deformation and damage conditions of the structure. This method of multi-modal data fusion can more comprehensively reflect the physical phenomena during the collision process, enhancing the reliability and accuracy of the remaining performance assessment of the protection facility.

[0012] In some embodiments, after the step of extracting collision feature parameters during the collision process from the accident video data through computer vision algorithms, the following steps are further included: When there are occlusions or angle limitations in the accident video data, identify the missing parameters in the collision feature parameters according to a preset collision parameter integrity detection mechanism; Based on the collision feature parameters and the final deformation state, construct a physical constraint equation set that conforms to the dynamic relationship of the collision process; Based on the physical constraint equation set, establish a parameter sensitivity matrix, which is used to quantify the influence degree of each collision feature parameter on the deformation result of the protection facility; Determine the key missing parameters that have the greatest impact on the digital twin model from the missing parameters according to the parameter sensitivity matrix; Randomly generate the key missing parameters according to the collision scene constraint conditions, and the collision scene constraint conditions include on-site environment, road specifications, and vehicle parameter information.

[0013] Through the above embodiments, the system addresses the problems of occlusion or angle limitations in common video data in actual scenarios. By establishing a parameter sensitivity matrix and a system of physical constraint equations, it can identify and supplement key missing parameters. This parameter correction method based on a physical model ensures reliable performance evaluation even when the data is incomplete, improving the practicality and adaptability of the evaluation method and enabling it to handle various complex accident scenarios.

[0014] In some embodiments, the step of randomly generating the key missing parameters according to the collision scenario constraint conditions specifically includes: Using the Monte Carlo method to generate multiple sets of parameter sampling values that conform to the collision scenario constraint conditions; Substituting each set of the parameter sampling values into the digital twin model for collision simulation to obtain the corresponding simulated deformation results; Screening out the optimal parameter combination corresponding to the simulated deformation result with the highest similarity to the actual deformation state through a similarity evaluation function; Determining the key missing parameters based on the optimal parameter combination.

[0015] Through the above embodiments, the system uses the Monte Carlo method to generate multiple sets of parameter sampling values and screens the optimal parameter combination through similarity evaluation, achieving accurate estimation of the missing parameters. This intelligent parameter optimization method not only ensures the rationality of physical constraints but also achieves the best match with the actual deformation state, improving the evaluation accuracy in the case of data loss and providing a reliable guarantee for rapid evaluation.

[0016] In some embodiments, after the step of acquiring the accident video data and vehicle parameter information uploaded by the accident vehicle, it further includes: Acquiring multi-source video data, where the multi-source video data includes the full-angle video of the post-accident protection facilities taken by on-site management personnel, the video recorded by surrounding monitoring devices during the accident, and the video uploaded by the driving recorders of passing vehicles; Based on a spatio-temporal alignment algorithm, matching and fusing the multi-source video data with the accident video data uploaded by the accident vehicle to obtain updated accident video data.

[0017] Through the above embodiments, the system significantly improves the integrity and reliability of accident data by integrating multi-source video data, including videos taken by on-site management personnel, surrounding monitoring devices, and passing vehicles, and performing spatio-temporal alignment and fusion. This all-round data acquisition method makes up for the limitations of a single perspective, provides more comprehensive information support for the performance evaluation of protection facilities, and improves the accuracy of evaluation results.

[0018] In some embodiments, before the step of simulating the collision occurrence process in the digital twin model and obtaining the stress distribution data and plastic deformation data of the protective facility, the following steps are further included: Obtain historical collision data corresponding to the collision location, where the historical collision data includes the time stamp of the historical collision event, collision characteristic parameters, and historical stress distribution data; Analyze the historical stress distribution data based on the material cumulative damage model and calculate the material fatigue degree of each key part; Modify the material parameters of the protective facility finite element model based on the material fatigue degree.

[0019] Through the above embodiments, the system calculates the material fatigue degree of the key parts by analyzing the historical collision data and the material cumulative damage model, and modifies the material parameters of the finite element model accordingly, so that the evaluation process takes into account the use history and cumulative damage influence of the protective facility. This evaluation method considering historical factors is more in line with the actual service state of the protective facility, and improves the accuracy and reliability of the remaining performance evaluation.

[0020] In a second aspect, the present application provides a traffic protection facility performance evaluation system, where the traffic protection facility performance evaluation system includes: one or more processors and a memory; The memory is coupled to the one or more processors, and the memory is used to store computer program code, where the computer program code includes computer instructions, and the one or more processors call the computer instructions so that the traffic protection facility performance evaluation system can implement a method for evaluating the remaining performance after an accident of a traffic protection facility based on digital twin provided in the above embodiments, which will not be elaborated here.

[0021] In a third aspect, the present application provides a computer-readable storage medium, including instructions, when the instructions run on the traffic protection facility performance evaluation system, the traffic protection facility performance evaluation system can implement a method for evaluating the remaining performance after an accident of a traffic protection facility based on digital twin provided in the above embodiments, which will not be elaborated here.

[0022] In a fourth aspect, the present application provides a computer program product, when the computer program product runs on the traffic protection facility performance evaluation system, the traffic protection facility performance evaluation system can implement a method for evaluating the remaining performance after an accident of a traffic protection facility based on digital twin provided in the above embodiments, which will not be elaborated here.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By obtaining accident video data and automatically extracting collision feature parameters in combination with computer vision technology, a digital twin model is established for simulation analysis, realizing the rapid and automated evaluation of the remaining performance of protective facilities. This method breaks through the limitations of traditional manual on-site detection. It can not only quickly obtain stress distribution and deformation data, but also deeply analyze the internal structural state of protective facilities, improving the evaluation efficiency and accuracy.

[0024] 2. Visual data, acoustic signals and multi-source videos are fused and analyzed, and the data missing situation is handled through the parameter sensitivity matrix and physical constraint equations. This method of multi-modal data fusion combined with physical model constraints not only improves the robustness of the evaluation, but also can accurately infer key parameters through intelligent algorithms under adverse conditions such as occlusion or limited viewing angle, ensuring the reliability of the evaluation results.

[0025] 3. By analyzing historical collision data and combining with the material cumulative damage model, the fatigue degree of protective facilities is quantitatively evaluated, and the material parameters of the digital twin model are dynamically adjusted accordingly. This dynamic evaluation method incorporates the usage history and cumulative damage of protective facilities into the evaluation system for the first time, making the evaluation results more in line with the actual service status and greatly improving the accuracy of the remaining performance evaluation. Description of the Drawings

[0026] Figure 1 is a schematic flow chart of a method for evaluating the remaining performance of traffic protective facilities after an accident based on digital twin in an embodiment of the present application; Figure 2 is another schematic flow chart of a method for evaluating the remaining performance of traffic protective facilities after an accident based on digital twin in an embodiment of the present application; Figure 3 is a schematic structural diagram of an entity device of a traffic protective facility performance evaluation system in an embodiment of the present application. Detailed Embodiments

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular forms "a", "an", "the", "above-mentioned", "said" and "this" are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations including one or more of the listed items.

[0028] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and should not be construed as implying or suggesting relative importance or implicitly indicating the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of this application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0029] It should be noted that for the convenience of description, the "Traffic Protection Facility Performance Evaluation System" in this application may also be abbreviated as the "System" in the following embodiments, and this abbreviation should not be a limitation of this application.

[0030] For ease of understanding, the method provided in this embodiment will be described in terms of its process below. Please refer to Figure 1 , which is a schematic flowchart of a method for evaluating the remaining performance of traffic protection facilities after an accident based on digital twins in the embodiments of this application.

[0031] S101. Obtain the accident video data and vehicle parameter information uploaded by the accident vehicle.

[0032] Among them, the accident vehicle refers to the vehicle that has a collision accident with the road traffic protection facilities; the accident video data refers to the video data uploaded by the in-vehicle monitoring system (such as a driving recorder) of the accident vehicle, which contains the complete process before, during, and after the accident. The vehicle parameter information represents the parameters related to the accident vehicle, including but not limited to vehicle mass, vehicle size, and vehicle stiffness.

[0033] Specifically, when the traffic protection facilities encounter a collision accident, the accident vehicle will upload the video data recording the whole process of the accident and its own vehicle parameter information. The traffic protection facility performance evaluation system receives these data to provide basic data support for subsequent operations such as analyzing collision characteristics and establishing a digital twin model.

[0034] Optionally, a data upload module is pre-installed at the time of vehicle factory. This module is connected to the vehicle's driving recorder and sensors. When an accident occurs, the driving recorder automatically stores the accident video data and obtains parameter information such as vehicle mass, size, and stiffness through the vehicle sensors. The data upload module uses the in-vehicle network, such as 4G or 5G network, to upload this data to the server address specified by the traffic protection facility performance evaluation system, and the system obtains the data from the server through a preset interface.

[0035] S102. Extract the collision feature parameters of the collision process from the accident video data through computer vision algorithms.

[0036] Specifically, the traffic protection facility performance evaluation system calls computer vision algorithms to process the acquired accident video data. By analyzing information such as the motion states and relative positions of vehicles and protection facilities in the video through algorithms, key parameters such as collision speed, collision angle, collision position, and deformation amount of the protection facilities are accurately extracted.

[0037] Optionally, the system first extracts a video frame sequence from the accident video data, processes each frame of the image using an image enhancement algorithm to improve the clarity and contrast of the image. Then, an object segmentation algorithm is used to separate the protection facilities and vehicles from the background to obtain the image of the protection facility area. Next, feature point detection and tracking are performed on the image of the protection facility area to obtain the motion data of the feature points. Finally, based on this motion data of the feature points, kinematic analysis methods are used to calculate the collision speed and collision angle, and 3D reconstruction methods are used to determine the collision position and the deformation amount of the protection facilities.

[0038] Optionally, the system can also preprocess the accident video data to remove noise and interference information first. Then, deep learning object detection models such as YOLO, Faster R-CNN, etc. are used to identify the vehicles and protection facilities in the video and determine their positions and contours. By analyzing the position changes of the vehicles and protection facilities in consecutive video frames, the collision speed and collision angle are calculated. Using stereo vision technology and combining video data captured by multiple cameras, 3D positioning of the collision position is performed, and at the same time, the deformation amount of the protection facilities is calculated through an image deformation analysis algorithm.

[0039] S103. Establish a digital twin model under accident conditions based on vehicle parameter information and collision characteristic parameters.

[0040] Specifically, the traffic protection facility performance evaluation system constructs a vehicle kinematic model and a protection facility finite element model respectively according to the acquired vehicle parameter information and collision characteristic parameters, and then combines them to form a digital twin model under accident conditions. When constructing the vehicle kinematic model, parameters such as vehicle mass, collision speed, and collision angle are used as inputs to determine the motion trajectory and state changes of the vehicle during the collision process. For the protection facility finite element model, based on parameters such as the structural design, material properties, collision position, and deformation amount of the protection facility, mesh division and mechanical parameter setting are performed on the protection facility to simulate its mechanical behavior under the action of collision forces. This digital twin model can truly reflect the interaction between vehicles and protection facilities under accident conditions and provide a basis for subsequent simulation of the collision process, obtaining stress distribution, and plastic deformation data.

[0041] It should be noted that the digital twin model is a virtual model corresponding to a real physical entity or system, and is used in this step to simulate the collision process between a vehicle and traffic protection facilities under accident conditions. It includes a vehicle kinematic model and a finite element model of the protection facilities. Among them, the vehicle kinematic model is used to describe the motion state of the vehicle during the collision process, such as the variation laws of speed, acceleration, displacement, etc. over time, and is constructed based on the vehicle parameter information and the speed, angle, etc. in the collision characteristic parameters; the finite element model of the protection facilities discretizes the protection facilities into a finite number of elements, and simulates its mechanical response during the collision process through numerical calculation methods, including stress, strain distribution and deformation conditions, and is constructed according to the structural characteristics, material properties of the protection facilities and the collision position, deformation amount, etc. in the collision characteristic parameters.

[0042] S104. Simulate the collision occurrence process in the digital twin model to obtain the stress distribution data and plastic deformation data of the protection facilities.

[0043] Specifically, the traffic protection facility performance evaluation system inputs the previously obtained vehicle parameter information, collision characteristic parameters, etc. into the digital twin model. In the model, the vehicle kinematic model simulates the motion trajectory and state changes of the vehicle during the collision according to the set parameters, and the generated collision force is transmitted to the finite element model of the protection facilities. The finite element model of the protection facilities, based on its own structure, material properties and the applied collision force, simulates the mechanical response of each element during the collision through numerical calculation methods. During the simulation process, the system records the stress change conditions of each part of the protection facilities to form stress distribution data; at the same time, it tracks information such as the area and deformation amount where the protection facilities undergo plastic deformation to obtain plastic deformation data.

[0044] Optionally, the system can use professional engineering simulation software, such as ANSYS. First, import the digital twin model constructed in step S103 into the software, including the vehicle kinematic model and the finite element model of the protection facilities. Set the relevant parameters for the collision simulation according to the accident video data, such as the collision duration, time step, etc. Then start the simulation calculation, and the software will perform the collision simulation according to the set parameters and the mechanical principles of the model. After the simulation is completed, through the post-processing module of the software, extract the stress distribution cloud map and plastic deformation data of the protection facilities, and the stress magnitude and plastic deformation degree of different parts can be visually viewed, and the relevant data files can be exported for subsequent analysis.

[0045] In addition, when the traffic protection facility performance evaluation system is about to conduct a new collision simulation, in order to consider the damage accumulated by the protection facility in past collisions, historical collision data corresponding to the collision location can be obtained before the collision simulation. Specifically, it can be retrieved from the database storing historical data. The database stores historical collision event records corresponding to each collision location. The system filters out the historical collision data corresponding to this location by precisely matching the coordinates or identification information of the collision location. These data are stored in a structured form, including timestamp, collision characteristic parameters, historical stress distribution data, etc.

[0046] The system inputs the historical stress distribution data into a pre-set material cumulative damage model (such as professional finite element analysis software like ANSYS). This model analyzes the stress history of each key part based on factors such as stress amplitude and number of loadings. For each key part, the model calculates the damage generated by each stress loading and accumulates these damages. For example, using the Miner linear cumulative damage theory, which believes that the fatigue damage of materials is the sum of the damages generated by each stress cycle. When the accumulated damage reaches 1, the material undergoes fatigue failure. The system substitutes the number of stress cycles and the corresponding stress amplitudes of each key part in the historical collision into the model for calculation, and finally obtains the material fatigue degrees of each key part. Then, based on the calculated material fatigue degrees of each key part, the system adjusts the material parameters in the finite element model of the protection facility. For example, for key parts with higher fatigue degrees, a certain proportion of their elastic modulus is reduced, and the yield strength is also correspondingly decreased, so that the finite element model better conforms to the actual mechanical properties of the protection facility after experiencing multiple collisions.

[0047] S105. Based on the stress distribution data and plastic deformation data, combined with the performance requirements in the protection facility design specifications, establish a remaining performance evaluation index system.

[0048] Among them, the protection facility design specifications refer to a series of standards and specifications followed during the design and construction of protection facilities, which stipulate various performance indicators and technical requirements that the protection facilities should possess; the remaining performance evaluation index system is a set of indicators used to evaluate the remaining performance state of traffic protection facilities after experiencing collision accidents, including structural strength, geometric dimensions, and functional integrity, etc.

[0049] Specifically, the traffic protection facility performance evaluation system first determines whether the stress at each part of the protection facility exceeds the allowable stress range specified in the design code based on the stress distribution data, so as to evaluate the damage condition in terms of structural strength. Then, according to the plastic deformation data, the geometric dimensions of the protection facility after deformation are measured and compared with the original dimensions in the design code to determine the degree of change in geometric dimensions. At the same time, based on considering the influence of stress distribution and plastic deformation on the function of the protection facility, its functional integrity is judged, such as whether the protection facility can still effectively block vehicles, guide traffic, etc. These evaluation results are integrated to form a remaining performance evaluation index system including aspects such as structural strength, geometric dimensions, and functional integrity.

[0050] Optionally, the system can adopt the Analytic Hierarchy Process (AHP). First, it is determined that the target layer of the remaining performance evaluation index system is the remaining performance evaluation of the protection facility, and the criterion layers are structural strength, geometric dimensions, and functional integrity. Then, for each criterion layer index, a judgment matrix is constructed according to the stress distribution data, plastic deformation data, and design code requirements. For example, for structural strength, the relative magnitudes of the impacts of stresses exceeding the allowable stress at different parts on the overall structural strength are compared. By calculating the eigenvector of the judgment matrix and performing a consistency check, the weights of each index are determined. Finally, the evaluation values of each index are comprehensively calculated according to the weights to obtain a preliminary result of the remaining performance evaluation of the protection facility, and further improve the remaining performance evaluation index system.

[0051] S106. Calculate the remaining performance index of the traffic protection facility according to the remaining performance evaluation index system, and match the corresponding maintenance and reinforcement suggestions based on the remaining performance index.

[0052] Specifically, the traffic protection facility performance evaluation system calculates the remaining performance index of the traffic protection facility by using a specific calculation method (such as the weighted average method, etc.) according to the weights of each index in the remaining performance evaluation index system and the corresponding evaluation values. For example, the weight of the structural strength index is w1, and the evaluation value is x1; the weight of the geometric dimensions index is w2, and the evaluation value is x2; the weight of the functional integrity index is w3, and the evaluation value is x3, then the remaining performance index I = w1 * x1 + w2 * x2 + w3 * x3.

[0053] After obtaining the remaining performance index, the system compares the remaining performance index with a preset threshold range. If the remaining performance index is higher than the threshold range, it indicates that the protection facility is less damaged and may only need simple local repair; if the remaining performance index is within the threshold range, it is determined that medium-level maintenance and reinforcement may be required; if the remaining performance index is lower than the threshold range, it is determined that the protection facility may need to be replaced as a whole. The system matches the corresponding maintenance and reinforcement suggestions based on these comparison results.

[0054] In the above embodiments, the system obtains accident video data and vehicle parameters, combines computer vision to extract collision features, and establishes a digital twin model for simulation analysis, so as to quickly obtain the stress distribution and deformation data of the protection facilities without on-site manual detection. Based on these data, a remaining performance evaluation index system is established, which can comprehensively evaluate the structural strength, geometric dimensions and functional integrity of the protection facilities, and give corresponding maintenance suggestions. This method avoids the subjectivity and time delay of traditional manual detection, can quickly and accurately evaluate the remaining performance state of the protection facilities after an accident, and improves the evaluation efficiency and accuracy.

[0055] For ease of understanding, the following further describes the process of the system in this embodiment extracting the collision feature parameters in the collision process from the accident video data through computer vision algorithms. Please refer to Figure 2 , which is another process schematic diagram of a method for evaluating the remaining performance of traffic protection facilities after an accident based on digital twins in the embodiments of the present application.

[0056] S201. Extract a video frame sequence from the accident video data, perform image enhancement and target segmentation processing, and obtain an image of the protection facility area.

[0057] Optionally, the system can use the video reading function of OpenCV to read the accident video data at a fixed frame rate, decompose the video into a series of image frames, and generate a video frame sequence. Then, call the image enhancement function of OpenCV, such as the CLAHE (Contrast Limited Adaptive Histogram Equalization) function, to perform enhancement processing on each frame of the image to improve the contrast and clarity of the image. Finally, use the target segmentation method based on contour detection. First, find the edge information in the image through an edge detection algorithm (such as the Canny edge detection), and then filter out the contour of the protection facility according to the shape characteristics and position information of the protection facility, and segment the protection facility from the background to obtain an image of the protection facility area.

[0058] S202. Detect and track feature points in the image of the protection facility area to obtain feature point motion data.

[0059] Specifically, the traffic protection facility performance evaluation system uses a specific feature point detection algorithm, such as the SIFT (Scale-Invariant Feature Transform) algorithm, to find points with unique features in the image of the protection facility area. These points can represent the local features of the protection facility. Then, use a feature point tracking algorithm, such as the KLT (Kanade - Lucas - Tomasi) tracking algorithm, to track the detected feature points in consecutive image frames. By continuously updating the position of the feature points in each frame of the image and recording their motion trajectories, the feature point motion data can be obtained.

[0060] S203. Extract the audio data from the accident video data, perform noise reduction and segmentation processing to obtain the collision acoustic signal.

[0061] Specifically, the traffic protection facility performance evaluation system can separate the audio data from the accident video data through a video processing tool. Then, a noise reduction algorithm, such as the noise reduction method based on wavelet transform, is used to process the audio data to remove the noise components therein and make the audio clearer. Next, according to the time information of the collision occurrence and the characteristics of the audio signal, the denoised audio is segmented. For example, taking the moment of collision as the center, audio segments of a certain duration are intercepted forward and backward to obtain the collision acoustic signal.

[0062] Optionally, the system can use professional audio processing software, such as Audacity. First, import the accident video data into the Audacity software, and the software will automatically separate the audio track. Then, use the built-in noise reduction plugin of Audacity to perform noise reduction processing on the audio by sampling the noise sample and setting appropriate noise reduction parameters. Then, according to the analysis of the video image, determine the time range of the collision occurrence, and use the marking and cutting tools in Audacity to segment the denoised audio according to the collision time range, and export to obtain the collision acoustic signal. It can be understood that other methods can also be used, such as using the audio processing toolbox of MATLAB, etc., which is not limited here.

[0063] S204. Perform time-frequency analysis on the collision acoustic signal to obtain a spectrogram, and then extract acoustic feature parameters from the spectrogram.

[0064] Specifically, the traffic protection facility performance evaluation system uses a time-frequency analysis algorithm, such as the short-time Fourier transform (STFT), to process the collision acoustic signal. STFT divides the audio signal into multiple short-time segments, performs Fourier transform on each segment, so as to obtain the frequency components of the signal at different time points and generate a spectrogram. Then, from the generated spectrogram, acoustic feature parameters are extracted through specific algorithms. For example, by finding the frequency corresponding to the energy peak in the spectrogram, the peak frequency of the collision sound wave is determined; by integrating the energy in different frequency bands in the spectrogram, the energy distribution is obtained. These acoustic feature parameters will be used for the subsequent optimization of the collision feature parameters.

[0065] S205. Filter and optimize the feature point motion data based on the acoustic feature parameters to obtain the optimized feature point motion data.

[0066] Specifically, the traffic protection facility performance evaluation system analyzes the motion data of feature points based on the collision information reflected by the acoustic feature parameters. For example, if the peak frequency of the collision sound wave is relatively high, it indicates that the collision is relatively severe, which may cause large fluctuations in the motion data of feature points. At this time, according to parameters such as energy distribution, the data points with large fluctuations can be smoothed. By establishing a mathematical model, the acoustic feature parameters are combined with the motion data of feature points to perform a filtering operation on the motion data of feature points, removing outliers and noise interference, and obtaining more accurate optimized motion data of feature points.

[0067] Optionally, the system can adopt the Kalman filtering algorithm. First, the acoustic feature parameters are used as auxiliary observation data, combined with the initial state of the motion data of feature points (such as initial position and velocity), to establish a Kalman filtering model. In the model, parameters such as the noise covariance matrix in the prediction and update processes are adjusted according to the acoustic feature parameters. For example, when the energy distribution of the collision sound wave shows that the collision is relatively complex, the measurement noise covariance is appropriately increased to more reasonably fuse the observation data. Then, through the prediction and update steps of the Kalman filter, the motion data of feature points are iteratively optimized to obtain optimized motion data of feature points.

[0068] S206: Based on the motion data of feature points, use the kinematic analysis method to calculate the collision speed and collision angle, and use the three-dimensional reconstruction method to determine the collision position and the deformation amount of the protection facility.

[0069] Specifically, the traffic protection facility performance evaluation system first calculates the collision speed and collision angle based on the motion data of feature points by using the kinematic analysis method. By analyzing the displacement change and time interval of feature points in consecutive image frames, the motion speed of feature points is calculated according to the definition of speed (speed equals displacement divided by time). Then, combined with the geometric relationship between the vehicle and the protection facility and the position information of the feature points on the vehicle or the protection facility, the collision speed of the accident vehicle is deduced. For the collision angle, by analyzing the direction change of the feature point motion trajectory and the initial direction of the protection facility, the collision angle is calculated by using mathematical methods such as trigonometric functions.

[0070] In determining the collision position and the deformation amount of the protection facility, the system adopts the three-dimensional reconstruction method. If there is multi-source video data (such as videos taken by on-site management personnel, videos recorded by surrounding monitoring devices, and videos uploaded by the driving recorders of passing vehicles), the system can use the feature point information in these videos from different perspectives. Through principles such as triangulation, three-dimensional models of the protection facility and the vehicle are constructed to accurately determine the collision position. For the deformation amount of the protection facility, the deformation amount of each part is calculated by comparing the shape differences of the three-dimensional model of the protection facility before and after the collision.

[0071] When only monocular video data is available, the system can also extract video frames from the accident video first. After image enhancement processing, the feature point motion trajectory can be obtained using feature point detection and tracking algorithms. Based on the assumption that the initial shape of the protection facility is known and the vehicle motion is in a plane, an improved structured light three-dimensional reconstruction algorithm is used to estimate the depth of feature points according to information such as image texture and edges, and calculate their three-dimensional coordinates in combination with camera parameters. By screening and analyzing the feature points related to the collision and combining with the three-dimensional model of the protection facility, the collision position is determined. The reconstructed three-dimensional model of the protection facility is compared with the initial model, and the deformation amount is obtained by calculating the distance difference of corresponding points through point cloud registration. Finally, the results are verified by comparing with physical laws and actual situations, and the assumptions and algorithms are adjusted and optimized, so as to complete the inference of three-dimensional information from two-dimensional video images, determine the collision position and the deformation amount of the protection facility.

[0072] In the above embodiment, the system optimizes the visual feature point motion data by introducing collision acoustic signal analysis, effectively improving the accuracy of collision feature parameters. Acoustic features can capture collision details that are difficult to obtain visually, such as internal structure deformation and damage conditions. This method of multi-modal data fusion can more comprehensively reflect the physical phenomena during the collision process, improving the reliability and accuracy of the remaining performance assessment of the protection facility.

[0073] S207. When there are occlusions or angle limitations in the accident video data, identify the missing parameters in the collision feature parameters according to the preset collision parameter integrity detection mechanism.

[0074] Specifically, the traffic protection facility performance evaluation system comprehensively checks the accident video data to determine whether there are occlusions or angle limitations. If so, the system activates the preset collision parameter integrity detection mechanism. This mechanism analyzes the extracted collision feature parameters according to the preset rules. For example, by comparing the motion of feature points in different video frames, if a feature point disappears or its motion trajectory is abnormal in some key frames, combined with the logical judgment of the collision process, it is determined which collision feature parameters may be missing. For example, if the contact position at the moment of collision between the vehicle and the protection facility cannot be clearly seen, then the collision position parameter may be missing; if the vehicle is partially occluded and its motion trajectory cannot be accurately tracked, the collision speed and collision angle parameters may be inaccurate or missing.

[0075] S208. Based on the collision feature parameters and the final deformation state, construct a physical constraint equation set that conforms to the dynamic relationship of the collision process.

[0076] Specifically, the traffic protection facility performance evaluation system is based on the existing collision characteristic parameters, combined with the final deformation state of the protection facilities, and based on dynamic principles such as Newton's laws of motion and the law of conservation of momentum. For example, based on parameters such as collision speed, vehicle mass and stiffness of the protection facilities, the law of conservation of momentum is used to establish equations about the motion state of the vehicle and protection facilities after the collision; based on the deformation and material properties of the protection facilities, the principle of elastic mechanics is used to establish equations describing the stress-strain relationship inside the protection facilities.

[0077] Optionally, the system can also be automatically generated with the help of professional mechanical analysis software. The collision characteristic parameters and the final deformation state data of the protective facilities are input into professional mechanical analysis software such as ANSYS. The software automatically generates a set of physical constraint equations that conform to the dynamic relationship of the collision process based on the input data based on its own mechanical algorithms and models. The software will consider factors such as material properties and geometric shapes to ensure the accuracy and completeness of the set of equations. It is understandable that other methods can also be used to achieve this, such as using an open source mechanical calculation library to input data in accordance with its prescribed method and format to generate a set of equations, which is not limited here.

[0078] S209. Establish a parameter sensitivity matrix based on the physical constraint equations and then determine the key missing parameters that have the greatest impact on the digital twin model from the missing parameters.

[0079] Specifically, the traffic protection facility performance evaluation system calculates the degree of influence of a small change in each collision characteristic parameter on the deformation result of the protection facility by mathematically analyzing the physical constraint equations. For example, by taking partial derivatives of the physical constraint equations, the rate of change of the deformation of the protection facility when each parameter changes is obtained, and these rates of change are converted into a matrix form to obtain the parameter sensitivity matrix. Then, based on the parameter sensitivity matrix, the degree of influence of each missing parameter on the deformation result of the protection facility is compared. The greater the degree of influence of the missing parameter, the greater the impact on the results of the digital twin model simulating the collision process and evaluating the remaining performance of the protection facility, thereby determining the key missing parameters that have the greatest impact on the digital twin model.

[0080] S210. Generate multiple sets of parameter sampling values using the Monte Carlo method, and substitute them into the digital twin model for collision simulation to obtain corresponding simulated deformation results.

[0081] Specifically, according to the collision scenario constraint conditions, the traffic protection facility performance evaluation system uses the Monte Carlo method to randomly generate multiple groups of parameter sampling values within the allowable parameter value range. For example, for the missing collision speed parameter, a reasonable value range is determined based on factors such as vehicle type, road speed limit, and on-site environment, and then multiple speed values are randomly generated within this range. These parameter groups containing the sampling values of the key missing parameters are sequentially substituted into the digital twin model. The vehicle kinematic model in the model simulates the vehicle's motion trajectory and collision process according to the parameters, and the finite element model of the protection facility simulates the mechanical response of the protection facility under the action of the collision force, so as to obtain the simulated deformation results of the protection facility corresponding to each group of parameters.

[0082] S211. Screen out the optimal parameter combination corresponding to the simulated deformation result with the highest similarity to the actual deformation state through the similarity evaluation function, and determine the key missing parameters.

[0083] Specifically, the traffic protection facility performance evaluation system inputs each group of simulated deformation results and the actual deformation state into the similarity evaluation function for calculation. The similarity evaluation function calculates the similarity degree between the simulated deformation result and the actual deformation state in terms of shape, size, position, etc. according to a preset algorithm (such as Euclidean distance, cosine similarity, etc.), and obtains a similarity value. The system compares the similarity values corresponding to all simulated deformation results, finds the group of simulated deformation results with the largest value, and the corresponding parameter sampling value combination is the optimal parameter combination. Extract the value of the key missing parameter from the optimal parameter combination, and this value is the most reasonable value determined by the system for supplementing the missing parameter.

[0084] In the above embodiment, aiming at the problems of occlusion or angle limitation of common video data in the actual scenario, the system can identify and supplement the key missing parameters by establishing a parameter sensitivity matrix and a physical constraint equation set. The Monte Carlo method is used to generate multiple groups of parameter sampling values, and the optimal parameter combination is screened through similarity evaluation, realizing the accurate estimation of the missing parameters. This intelligent parameter optimization method not only ensures the rationality of physical constraints but also achieves the best match with the actual deformation state, improving the evaluation accuracy in the case of data missing.

[0085] The traffic protection facility performance evaluation system of the embodiment of the present invention is applied to an electronic device. Figure 3 The architecture diagram of the electronic device suitable for implementing the embodiment of the present invention is shown.

[0086] It should be noted that Figure 3 The shown electronic device is only an example and should not bring any limitation to the functions and usage scope of the embodiment of the present invention.

[0087] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructions (computer programs), or by controlling related hardware through instructions (computer programs). These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. The electronic device of this embodiment includes a storage medium and a processor. Among them, multiple instructions are stored in the storage medium, and these instructions can be loaded by the processor to execute any step of the method provided in the embodiments of the present invention.

[0088] Specifically, the storage medium and the processor are directly or indirectly electrically connected to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more signal lines. The computer execution instructions for implementing the data access control method are stored in the storage medium, including at least one software function module that can be stored in the storage medium in the form of software or firmware. The processor executes various functional applications and data processing by running the software programs and modules stored in the storage medium. The storage medium can be, but is not limited to, a random access storage medium (Random Access Memory, abbreviated as RAM), a read-only storage medium (Read Only Memory, abbreviated as ROM), a programmable read-only storage medium (Programmable Read-Only Memory, abbreviated as PROM), an erasable read-only storage medium (Erasable Programmable Read-Only Memory, abbreviated as EPROM), an electrically erasable read-only storage medium (Electric Erasable Programmable Read-Only Memory, abbreviated as EEPROM), etc. Among them, the storage medium is used to store programs, and the processor executes the programs after receiving the execution instructions.

[0089] Furthermore, the software programs and modules in the above storage medium may also include an operating system, which may include various software components and / or drivers for managing system tasks (such as memory management, storage device control, power management, etc.), and can communicate with various hardware or software components to provide a running environment for other software components. The processor can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor can be a general-purpose processor, including a central processing unit (Central Processing Unit, abbreviated as CPU), a network processor (Network Processor, abbreviated as NP), etc., which can implement or execute the various methods, steps, and logic flow block diagrams disclosed in this embodiment. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0090] Since the instructions stored in the storage medium can execute the steps in any of the methods provided by the embodiments of the present invention, the beneficial effects of any of the methods provided by the embodiments of the present invention can be achieved. For details, see the previous embodiments and will not be elaborated here.

[0091] As described above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A method for evaluating the remaining performance of traffic protection facilities after an accident based on digital twins, which is applied to a traffic protection facility performance evaluation system, and is characterized in that The method includes: Obtaining accident video data and vehicle parameter information uploaded by the accident vehicle. The accident video data includes the complete process videos before, during, and after the accident, and the vehicle parameter information includes vehicle mass, vehicle dimensions, and vehicle stiffness; Extracting collision characteristic parameters of the collision process from the accident video data through computer vision algorithms. The collision characteristic parameters include collision speed, collision angle, collision position, and deformation amount of the protection facility; Establishing a digital twin model under accident conditions based on the vehicle parameter information and the collision characteristic parameters. The digital twin model includes a vehicle kinematic model and a finite element model of the protection facility; Simulating the collision occurrence process in the digital twin model to obtain stress distribution data and plastic deformation data of the protection facility; Based on the stress distribution data and the plastic deformation data, combined with the performance requirements in the protection facility design specifications, establishing a remaining performance evaluation index system. The remaining performance evaluation index system includes structural strength, geometric dimensions, and functional integrity; According to the remaining performance evaluation index system, calculating the remaining performance index of the traffic protection facility, and matching the corresponding repair and reinforcement suggestions based on the remaining performance index.

2. The method according to claim 1, wherein The step of extracting collision characteristic parameters of the collision process from the accident video data through computer vision algorithms specifically includes: Extracting a video frame sequence from the accident video data and performing image enhancement and target segmentation processing to obtain an image of the protection facility area; Performing feature point detection and tracking on the image of the protection facility area to obtain feature point motion data; Based on the feature point motion data, using kinematic analysis methods to calculate the collision speed and collision angle, and using three-dimensional reconstruction methods to determine the collision position and the deformation amount of the protection facility.

3. The method according to claim 2, characterized in that, After the step of performing feature point detection and tracking on the image of the protection facility area to obtain feature point motion data, it further includes: Extracting audio data from the accident video data and performing noise reduction and segmentation processing to obtain a collision acoustic signal; Performing time-frequency analysis on the collision acoustic signal to obtain a spectrogram, and then extracting acoustic characteristic parameters from the spectrogram. The acoustic characteristic parameters include the peak frequency and energy distribution of the collision sound wave; Based on the acoustic characteristic parameters, filtering and optimizing the feature point motion data to obtain optimized feature point motion data.

4. The method according to claim 1, wherein After the step of extracting collision characteristic parameters of the collision process from the accident video data through computer vision algorithms, it further includes: When there are occlusions or angle limitations in the accident video data, identifying missing parameters in the collision characteristic parameters according to a preset collision parameter integrity detection mechanism; Based on the collision characteristic parameters and the final deformation state, constructing a physical constraint equation set that conforms to the dynamic relationship of the collision process; Based on the physical constraint equation set, establishing a parameter sensitivity matrix, which is used to quantify the influence degree of each collision characteristic parameter on the deformation result of the protection facility; Determining the key missing parameters that have the greatest impact on the digital twin model from the missing parameters according to the parameter sensitivity matrix; Randomly generate the key missing parameters according to the collision scenario constraint conditions, where the collision scenario constraint conditions include the on-site environment, road specifications, and vehicle parameter information.

5. The method according to claim 4, wherein The step of randomly generating the key missing parameters according to the collision scenario constraint conditions specifically includes: Adopt the Monte Carlo method to generate multiple groups of parameter sampling values, and the parameter sampling values conform to the collision scenario constraint conditions; Substitute each group of the parameter sampling values into the digital twin model for collision simulation to obtain the corresponding simulated deformation results; Screen out the optimal parameter combination corresponding to the simulated deformation result with the highest similarity to the actual deformation state through the similarity evaluation function; Determine the key missing parameters based on the optimal parameter combination.

6. The method according to claim 1, characterized in that, After the step of obtaining the accident video data and vehicle parameter information uploaded by the accident vehicle, it further includes: Obtain multi-source video data, where the multi-source video data includes the full-angle video of the post-accident protection facilities taken by on-site management personnel, the video recorded by surrounding monitoring devices during the accident, and the video uploaded by the driving recorders of passing vehicles; Based on the spatio-temporal alignment algorithm, match and fuse the multi-source video data with the accident video data uploaded by the accident vehicle to obtain the updated accident video data.

7. The method according to claim 1, characterized in that, Before the step of simulating the collision occurrence process in the digital twin model to obtain the stress distribution data and plastic deformation data of the protection facilities, it further includes: Obtain the historical collision data corresponding to the collision location, where the historical collision data includes the time stamp of the historical collision event, the collision characteristic parameters, and the historical stress distribution data; Analyze the historical stress distribution data according to the material cumulative damage model and calculate the material fatigue degree of each key part; Based on the material fatigue degree, correct the material parameters of the protection facility finite element model.

8. A traffic protection facility performance evaluation system, characterized in that The traffic protection facility performance evaluation system includes: one or more processors and a memory; The memory is coupled to the one or more processors, and the memory is used to store computer program code, where the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the traffic protection facility performance evaluation system to execute the method according to any one of claims 1-7.

9. A computer-readable storage medium, comprising instructions, characterized in that, When the instruction runs on the traffic protection facility performance evaluation system, it enables the traffic protection facility performance evaluation system to execute the method according to any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product runs on the traffic protection facility performance evaluation system, it enables the traffic protection facility performance evaluation system to execute the method according to any one of claims 1-7.

Images