533 results about "Traffic crash" patented technology

The invention discloses a road traffic accident cause analysis and responsibility judgment method and system, and relates to the technical field of traffic management data processing, and the method comprises the steps: carrying out the credible collection of multi-source data, and constructing an evidence chain; performing multi-source data preprocessing and cross-modal fusion optimization; performing multi-dimensional cause intelligent analysis; performing responsibility judgment based on a quantification rule; and the responsibility judgment result is subjected to multi-dimensional rechecking and rule iteration adaptation. According to the road traffic accident cause analysis and responsibility judgment method and system, through a full-process closed-loop design of data acquisition, preprocessing, cause analysis, responsibility judgment, re-checking iteration and report evidence storage, technologies such as multi-source perception and an AI algorithm are integrated; the method solves the problems of uncredible evidence chain, one-sided cause traceability, non-uniform judgment standard, low cooperation efficiency and the like in traditional accident processing, realizes intelligence, standardization, compliance and traceability of accident processing, remarkably improves the credibility, processing efficiency and judicial suitability of a judgment result, and provides core technical support for modernization of traffic control.

The invention relates to a traffic accident intelligent detection system and method based on a YOLOv12 improved architecture. The system comprises a YOLOv12 enhanced feature extraction network, a multi-scale detection head, a time sequence information fusion module, a real-time reasoning optimization engine and an intelligent decision fusion system, and realizes collaborative optimization of local feature enhancement and global context modeling by constructing six core technology modules and adopting collaborative learning of a C2f-Attention mechanism and deformable convolution. According to the method, a composite loss function special for traffic accidents is innovatively designed, and adaptive fusion of multi-scale features and difficult sample mining are realized through a multi-objective optimization mechanism of Enhanced Focus Loss, IoU-aware Loss and Severage-aware Loss. According to the method, the problems of low detection precision and false alarm and missing alarm caused by illumination variation, shielding and motion blur in a traffic monitoring scene are effectively solved, in the test of an AccidentsDesection YOLOv8 data set, the mAP at 0.5 reaches 91.27% and is improved by 8.6% compared with that of YOLOv8, the reasoning speed reaches 67 FPS, experimental results show that the system has excellent performance in the aspects of detection precision, real-time performance and model compression, and the method is suitable for popularization and application. The method achieves a remarkable effect in traffic accident intelligent identification, and has a remarkable technical effect and industrial application value.

The present application belongs to the technical field of highway traffic control. Disclosed in the present application are a method and device for dynamically evaluating the degree of impact of a highway traffic accident, which method and device are used for solving the existing technical problem of after a traffic accident occurs on a highway, transportation efficiency being prone to being subjected to severe negative impacts caused by it being difficult to dynamically predict the degree of impact of the highway traffic accident, which is not conducive to effectively estimating the severity of the accident. The method comprises: determining a local highway network of an accident site; performing index fusion on preset structural indexes and functional indexes, and performing continuous equalization processing on the severity of a highway traffic accident on the basis of comprehensive indexes obtained after fusion, so as to obtain an accident severity evaluation system for the highway traffic accident; and performing data fusion on information of the natural evolution patterns of traffic flows, information of the severity of the accident at the moment when the accident occurs, and information of abrupt changes in the traffic flows that are induced by the accident, so as to obtain an accident impact factor system of the highway traffic accident.

The invention discloses a traffic drainage signal management method and system based on a Beidou system, and relates to the technical field of Beidou satellite navigation, and the method comprises the steps: obtaining the real-time traffic data of the position and speed of a vehicle at each intersection based on the Beidou system, and transmitting the real-time traffic data to a traffic management platform; carrying out feature matching on the accident identification model trained based on historical emergency data and a deep learning algorithm and real-time traffic data, and identifying an emergency; and constructing a real-time traffic situation map in combination with the identified emergency information and the traffic condition information of the intersections. The vehicle position and speed data are obtained in real time through the Beidou system, emergencies are rapidly recognized and signal lamp timing is dynamically adjusted in combination with a deep learning algorithm, and compared with a traditional fixed timing scheme, traffic flow changes can be responded in real time, frequent start and stop of vehicles are reduced, the road passing efficiency is improved, and the traffic safety is improved. And particularly, in a sudden traffic accident or construction scene, a signal strategy can be quickly optimized, and traffic congestion aggravation is avoided.

The invention discloses a visual angle robust traffic accident detection method based on space-time attention domain adaptation. The method comprises the following steps: constructing a double-flow network architecture; the progressive multi-granularity spatial domain adaptation module processes appearance change caused by a visual angle through global and local feature alignment, and uses SSAM to generate a semantic mask and entropy-guided mobility weight to realize accurate spatial domain adaptation; the time collaborative attention module realizes dynamic alignment of accident related time information between different visual angles through cross-domain video clip correlation calculation and a collaborative attention mechanism; a comprehensive contrast learning strategy is introduced, and noise robust motion feature learning is enhanced; a multi-view-angle joint training strategy is adopted, cross-domain knowledge migration is carried out by using monitoring view angle, vehicle-mounted view angle and unmanned aerial vehicle view angle data, and model optimization is realized through a comprehensive loss function. According to the method, the performance is remarkably improved in a cross-view migration task, and an efficient solution is provided for a unified multi-view traffic accident detection system.

The invention discloses a large transport intelligent line selection implementation method fusing a graph neural network and a large language model, and the method comprises the following steps: S1, constructing multi-source features to a unified high-dimensional vector space according to large transport historical data and real-time environment data; s2, constructing a structured prompt template through multi-modal input; s3, based on the structured prompt template, generating a candidate route set in combination with a large language model; s4, dynamically optimizing the candidate route set through a reinforcement learning framework, and outputting an optimal route. Compared with the prior art, the method has the advantages that intelligent generation and dynamic decision making of a large transportation route are realized through multi-modal feature fusion, graph structure semantic compression and large model reasoning optimization, and the transportation reliability, safety and economical efficiency can be remarkably improved; traffic accidents can be responded in real time to realize route adjustment, routes are dynamically optimized to reduce passage cost, planning quality is continuously improved through historical case learning, and safe and punctual delivery of overrun goods is guaranteed.

The invention relates to the technical field of intelligent traffic, in particular to a precise traffic accident early warning system based on multi-sensor fusion. Comprising a multi-modal data acquisition unit; the intelligent central processing unit is used for generating differentiated early warning instructions adaptive to different traffic participant types based on the three-dimensional risk decoupling result and the collision risk prediction result; a grading differentiation early warning execution unit; and an emergency linkage unit. According to the method, a meta-task sampling strategy is constructed through causal effect value grading and intersection topological feature clustering, edge lightweight adaptation is achieved in combination with gradient compression, parameter self-evolution is achieved through sliding window monitoring and knowledge distillation, and edge computing power and early warning precision are balanced. According to the invention, the multi-modal graded differentiated early warning terminals adapted to different traffic participants and a 5G edge low-delay public security and emergency linkage mechanism are matched, the early warning pertinence and the disposal practicability are quickly responded and improved, and the problem of insufficient edge adaptation and precision balance in the prior art is solved.

The invention relates to the technical field of artificial intelligence and intelligent traffic, in particular to a traffic accident liability intelligent affirmation method and system based on a multi-agent cooperation mechanism and a large language model. The whole process of credible information verification, responsibility affirmation reasoning and standard document generation is simulated. Wherein the legal expert agent adopts a competing mechanism, and the fairness and the accuracy of affirmation are improved through preliminary affirmation, double-party defense and final judgment. The reasoning ability of a large language model and accurate knowledge retrieval of a retrieval enhancement generation technology are fused, the real-time performance and accuracy of legal clause quotation are ensured, and a road traffic accident identification document conforming to specifications is automatically generated. The method effectively solves the problems that traditional manual identification is low in efficiency and high in subjectivity, and an existing intelligent method lacks interpretability and legal accuracy.

The invention provides a traffic accident emergency response method and system based on vehicle state perception, and relates to the technical field of intelligent traffic, and the method comprises the steps: deploying a vehicle-mounted sensor and roadside perception equipment, and collecting and obtaining vehicle state data and traffic environment data; constructing a traffic accident analysis index set, and analyzing the vehicle state data and the traffic environment data; building a traffic accident risk assessment model, assessing the traffic accident index parameter set, and outputting a traffic accident risk coefficient; triggering a target emergency response strategy; and analyzing the traffic accident index parameter set to obtain a target emergency strategy parameter, and carrying out accident emergency response processing. According to the method and the device, the technical problem of emergency response delay caused by inaccurate risk assessment of the traffic accident in a complex traffic situation in the prior art is solved, and the processing efficiency of the traffic accident is improved by comprehensively considering the multi-dimensional data, accurately assessing the risk of the traffic accident and automatically triggering the emergency strategy.

The invention relates to a traffic accident scene live-action reconstruction method and device, equipment and a medium, and the method comprises the steps: carrying out the thinning optimization of a traffic accident scene image through employing a preset Douglas-Peucker algorithm, so as to generate a thinned and optimized traffic accident scene image; performing patch expansion and filtering on the sparse point cloud by adopting a preset PMVS algorithm to obtain dense point cloud corresponding to each traffic accident element; taking the dense point clouds corresponding to the traffic accident elements as geometric constraints, taking the thinned and optimized traffic accident scene image as a texture data source to construct a training data set, and training a preset neural radiation field model to be converged by adopting the training data set to determine a traffic accident scene three-dimensional model; and inputting a to-be-reconstructed traffic accident scene image into the traffic accident scene three-dimensional model to extract three-dimensional space information corresponding to each traffic accident element. According to the method, the accident scene investigation and modeling period can be greatly shortened, and the road blocking time is greatly shortened.

The invention relates to the technical field of unmanned vehicles, and discloses an unmanned vehicle auxiliary lane changing decision-making method based on multi-modal fusion, and the method comprises the steps: synthesizing a visual image of a densely overlapped vehicle contour through a generative adversarial network, shielding missing laser point cloud, radar signals of clutter interference, and other analog data; samples are expanded in combination with oversampling and transfer learning, and the problem of insufficient sample size is solved. A dynamic attention mechanism is utilized, correlation modes of multi-modal features such as vision, laser and radar in the scene are mainly learned, the scene is quickly recognized through a long-tail scene adaptation module of a meta-learning framework, and decision parameters are dynamically adjusted. In training, a mixed loss function is adopted to reinforce learning, sudden plug triggering is re-evaluated during real-time decision making, and meanwhile, the model is continuously optimized by relying on a feedback iteration mechanism. Finally, the model can accurately identify a plugging scene, the safety space is judged by fusing multi-modal features, dynamic changes are dealt with, and the traffic accident risk caused by plugging is reduced.

The invention provides a smart city traffic abnormity monitoring method and system based on an Internet of Things large model, and relates to the field of Internet of Things and smart city traffic management. The system comprises an abnormity judgment module and a diversion module. The abnormity judgment module is configured to perform abnormity judgment on the target area according to the multi-source data and determine a plurality of hidden danger hot areas; the diversion module is configured to determine a plurality of standby routes according to the judgment result and the regional road network topological map; determining a plurality of main routes according to the plurality of standby routes and the position information of the plurality of variable information boards, generating a diversion instruction, and sending the diversion instruction to the emergency supervision object platform; and based on the diversion instruction, controlling a plurality of variable information boards to display a sketch of a corresponding main route, and controlling traffic lights on a plurality of standby routes to perform green light signal display according to a passing period. According to the method, the main pushing route of the variable information boards can be reasonably determined and controlled, potential traffic accident hidden dangers are prevented, and traffic safety is guaranteed.

The invention discloses an automatic driving traffic accident intelligent analysis system based on sand table simulation, and relates to the technical field of vehicle automatic driving. The system comprises a sand table simulation scene construction module which is used for simulating a high-risk scene of an automatic driving traffic accident and providing a physical and digital twinborn combined experiment environment; the automatic driving data acquisition and perception module is used for acquiring vehicle sensor data in real time and realizing environment perception and behavior decision simulation; the traffic accident intelligent analysis core module is used for realizing accident cause analysis and responsibility determination based on multi-source data and a knowledge base; and the cloud control platform and data visualization module is used for monitoring sand table simulation data in real time and providing a visual interface and an interaction function. The system can simulate and analyze the performance of automatic driving in various complex traffic environments and perform analysis and prediction, the analysis and prediction accuracy is high, and the analysis effect is better.

The invention relates to an expressway road condition monitoring system and method, and belongs to the field of expressway road condition monitoring, and the system comprises an ultrasonic sensor which is used for detecting whether branches grow into a guardrail of an expressway or not; the pressure sensor is arranged on the guardrail of the expressway and is used for detecting whether the guardrail of the expressway is bent or not; the image recognition module is used for detecting whether a traffic accident occurs in the target road section; the ESP32 single-chip microcomputer is connected with the temperature and humidity sensor, the ultrasonic sensor and the pressure sensor and gives an alarm when the branches grow into the guardrail of the expressway, whether the guardrail of the expressway is bent or not and a traffic accident happens to a target road section. According to the invention, through the strong operation processing capability and low power consumption characteristic of the ESP32, in combination with various sensors such as ultrasonic, pressure, temperature and humidity sensors of the image recognition module, complex road condition scenes can be accurately analyzed, and abnormal conditions such as traffic accidents can be timely captured, so that road condition monitoring is more comprehensive and meticulous.

The invention discloses a traffic abnormal event detection method based on multi-source data fusion, and the method comprises the steps: obtaining target detection data through a single millimeter wave radar and a plurality of industrial cameras according to the characteristics of a sensor; obtaining fusion information by associating the multi-source information; tracking the fusion target to form a target trajectory, and performing smooth filtering on the trajectory to form a filtered target trajectory; violation event detection is realized by analyzing a target track, wherein violation events comprise invasion, overspeed, underspeed, lane change, illegal parking, reverse driving and line pressing; traffic state event detection is achieved by analyzing the correlation between the lane-level detection area and target data, traffic state events include congestion, event time is aligned for event evidence leaving, and event videos and event photos are included; and for a specific event, event recheck is carried out through target track anti-shake verification, and event qualification and reporting are carried out. The traffic safety risk can be effectively and intelligently prevented and controlled, the traffic efficiency is optimized, and the traffic accident occurrence probability is reduced.

This application discloses a highway traffic diversion alarm broadcasting system and method based on AI recognition, relating to the field of highway diversion technology. This application analyzes the load of each vehicle and the driving experience, historical highway violations, and historical highway accidents of each driver to divert vehicles, diverting high-risk vehicles to safer highways, thereby ensuring vehicle safety. Different safe following distances and speeds are set for different highways to avoid traffic accidents caused by excessively close following distances or excessive speeds. Furthermore, the system monitors all vehicles traveling on each highway, broadcasting warnings when accidents or violations are detected, which enhances deterrence against drivers and prevents repeated violations from leading to traffic accidents.

The invention relates to the technical field of intelligent traffic and artificial intelligence crossing, and particularly discloses a traffic accident intelligent responsibility affirmation AI auxiliary method and system based on a large model, and the method comprises the steps: obtaining a responsibility judgment knowledge base, and processing the knowledge in the responsibility judgment knowledge base; key information of the traffic accident case is acquired, and occurrence passing information of the traffic accident case is generated according to the key information of the traffic accident case; through a large model, according to a responsibility determination knowledge base, the key information of the traffic accident case and the occurrence information of the traffic accident case, determining the wrong behavior and responsibility of the traffic accident party; and outputting error-passing behaviors and responsibilities of the traffic accident parties, and collecting actual responsibility affirmation results of the traffic accident parties to form complete feedback data of traffic accident cases. According to the traffic accident intelligent responsibility determination AI auxiliary method based on the large model provided by the invention, standardization, intelligence and high efficiency of a responsibility determination process can be realized.

The invention discloses a traffic safety desktop war game deduction system with a multimedia acquisition function, and the system comprises a desktop deduction module which is used for providing a traffic accident case background, role rules and task distribution, supporting the synchronous operation of multiple roles, such as traffic police, rescue personnel, wounded personnel and on-site command, and achieving the information sharing and task cooperation among the roles; modularizing the entity sand table model; according to the invention, the portability and scene flexibility are greatly improved: the modularized entity sand table adopts a light foldable base map and a magnetic attraction / buckle type assembly, the whole modularized entity sand table can be loaded into a common suitcase, a single person can complete cross-site carrying and deployment, and the problem of'immobilization 'of a traditional large simulation device is solved; and meanwhile, various accident scenes such as urban roads, expressways and rural lanes are supported to be quickly assembled, the multi-scene requirements of police school classroom teaching, basic-level traffic police actual combat training and the like are met, and the deployment efficiency is improved.

The invention discloses a traffic accident detection method based on a VAE model, and the method comprises the following steps: obtaining original data, and carrying out the primary processing of the obtained data; dividing a data set; converting a data format into a tensor format; performing cross validation on the data; pre-training the VAE model, adjusting and training a classifier, calculating a sample reconstruction error, KL divergence, a potential spatial distance and an output probability of the classifier, and calculating a mixed score through a mixed scoring formula; according to the method, the accuracy of a traffic accident detection model is improved, the accuracy of accident detection in an unbalanced data scene is effectively improved by fusing the characterization learning ability of the VAE and the discrimination ability of the classifier, and a mixed scoring mechanism is further introduced, so that the accuracy of the traffic accident detection in the unbalanced data scene is improved. Information of multiple dimensions such as VAE reconstruction error, KL divergence, potential spatial distance and classifier output probability is integrated, and the robustness of the model is enhanced.

The application discloses a risk road section prediction method and system, computer equipment and a storage medium, and relates to the technical field of road safety. The method comprises the following steps: acquiring street view image data of traffic accidents and accident-free road sections, inputting the street view image data of the traffic accidents and the accident-free road sections into a semantic segmentation model, outputting a predicted segmentation image, and obtaining the pixel proportion of different environmental elements of the road according to the segmented image; meanwhile, a risk road section prediction model is constructed based on an XGBOOST algorithm; the pixel proportions of different environmental elements are input into the risk road section prediction model, and the prediction probability of each street view image for an accident is output; the accuracy and stability of XGBOOST prediction are enhanced by deeply analyzing the pixel proportions of various objects in the visual environment of drivers, the risk of local transportation lines is accurately evaluated and analyzed in terms of environmental elements, and the method has a wide application prospect.

The embodiment of the invention discloses a processing method and device for an occluded aerial photo and electronic equipment. The method comprises the following steps: acquiring an aerial image of a road environment in which a traffic accident occurs, and acquiring natural language description information of a user for a shielding object in the aerial image; determining, by a visual language model, a first processed image based on the aerial image and the natural language description information; and based on the first processing image and the natural language description information, outputting a restored image corresponding to the aerial image by a generative adversarial network, the generative adversarial network comprising a dual-path discriminator for combined judgment by an image receiving path and a language description receiving path. In the embodiment, the labor cost of supervised learning of a traditional deep learning model is reduced, in the processing process, only the user needs to carry out brief natural language description, the user does not need to manually recognize and select a shielding part and define the shielding object, accurate processing of the shielding object can be completed, the working process is simplified, and the processing efficiency is improved.

The application discloses a vehicle braking method, device, equipment and storage medium. In the process of driving a target vehicle, if the state of a traffic signal at a road intersection changes to prohibition of passing after a first preset time period, a first distance between the target vehicle and a stop line of the intersection is determined. If the first distance is not less than a first preset threshold, a first braking force is determined to brake the target vehicle. If the first distance is less than the first preset threshold, it is determined whether the rear wheel of the target vehicle can cross the stop line of the intersection within the first preset time period without changing the current speed. If yes, a continue driving signal is sent to the target vehicle. If no, a second braking force is determined to brake the target vehicle. The scheme can brake or pass the intersection in time when the vehicle encounters a limit situation, such as the traffic signal changing to prohibition of passing immediately, effectively reduces the phenomenon of vehicle running a red light, and can also avoid traffic accidents caused by the vehicle not reacting in time.

The invention discloses a mountainous area expressway secondary accident active prevention and control method and system, and the method comprises the steps: carrying out the global monitoring of a mountainous area expressway continuous slope road section, extracting traffic scene features, and identifying an accident scene; in combination with the current traffic flow state, the road static characteristics, the environmental conditions and the vehicle behaviors, estimating the initial accident occurrence risk based on an accident risk estimation model of risk coupling; the method comprises the following steps: predicting whether a secondary accident of the continuous slope bending of the expressway occurs or not based on a Bayesian optimization random forest model, if so, predicting the occurrence time and place of the secondary accident, and then performing active prevention and control through a warning means and a forced control means. According to the invention, through air-ground integrated monitoring and intelligent control, rapid detection of a primary traffic accident and accurate assessment and active prevention control of a secondary accident risk are realized, so that the traffic accident frequency and damage degree of a high-risk road section are greatly reduced.

The road construction vehicle-mounted early warning system comprises a radar sensing module, a first communication module, a controller module and a warning module, the output end of the radar sensing module is electrically connected with the input end of the first communication module, and the radar sensing module is used for detecting the running speed and vehicle distance data of a vehicle; the output end of the first communication module is electrically connected with the input end of the controller module, and the first communication module is used for transmitting the detected vehicle running speed and vehicle distance data to the controller module to screen abnormal data; the output end of the controller module is electrically connected with the input end of the warning module, and the controller module is used for carrying out early warning and reminding on a vehicle corresponding to the abnormal data; the driving speed and the vehicle distance data of the vehicle are monitored in real time through the radar sensing module, the detected data are transmitted to the controller module to be processed, the warning module carries out acousto-optic reminding on the vehicle with abnormal data, the potential collision risk can be found and reminded in time for a driver, the occurrence rate of traffic accidents is reduced, and the safety is improved.

The invention discloses a traffic scene simulation method based on an accident report, and aims to solve the problem of insufficient complex dynamic scene data in traffic simulation. According to the method, physical clues (such as vehicle speed, collision angle, quality and the like) and context information (such as road types and environmental conditions) are extracted from a real traffic accident report, and a high-fidelity traffic scene data set is generated. In foreground processing, the extracted physical clues are used for accurately constructing a vehicle motion trajectory, diversified dynamic traffic scenes are generated through a pre-collision trajectory planning algorithm, and trajectory prediction is optimized in combination with a large language model to ensure that the real physical law is met. In background processing, high-quality three-dimensional scene reconstruction is carried out according to environment information in an accident report, a vivid background image is generated, and the consistency of a visual angle and illumination is kept. And finally, through fusion of the foreground vehicle track and the background image, a visual and physical double real traffic simulation video is generated. The method significantly improves the diversity and authenticity of traffic scenes, reduces the dependence on real data collection, is widely suitable for high-fidelity traffic simulation scenes such as automatic driving test and traffic flow analysis, and effectively promotes the progress of the traffic simulation technology.

The application provides a Gaussian ball clustering method for three-dimensional reconstruction of a highway traffic accident site, comprising the following steps: S1, collecting multi-view images of the accident site, solving camera parameters and initializing a three-dimensional Gaussian ball scene; S2, constructing a rendering training loop based on the initialized three-dimensional Gaussian ball scene, and adaptively adjusting the Gaussian ball density through cloning or splitting operations; S3, extracting the spatial and color features of the three-dimensional Gaussian ball, and constructing a multi-modal physical feature set; S4, calculating an adaptive discrete step length based on the multi-modal physical feature set, and mapping the continuous features into discrete coordinates; S5, generating a multi-modal hash index based on the discrete coordinates, and realizing continuous rearrangement of the same type of Gaussian ball in the display memory through GPU parallel sorting; S6, performing physical conservation formula parameter aggregation on the same type of Gaussian ball continuously stored in the display memory, releasing redundant display memory and simplifying the three-dimensional Gaussian ball set; and S7, completing all training iterations, and exporting the final simplified three-dimensional Gaussian ball set into a general PLY format file.

We will build a circular integrated service platform that provides a circular integrated service delivery system that enables the arrangement of replacement vehicles based on a personality score and the reduction of vehicle transportation costs when a non-life insurance policyholder is involved in a car accident. [Solution] The cyclical integrated service provision system 4 according to the present invention comprises a replacement vehicle service system 5 equipped with multiple replacement vehicle stations 15 located throughout the country or in certain regions, which provide replacement vehicles to insured persons with a replacement vehicle expense rider on their automobile insurance via a network. The insured person's member terminal 1 is equipped with a function for receiving replacement vehicle usage services based on a personality score. The management server 7 is characterized in that, when the accident reception center terminal 9 receives a traffic accident notification from the member terminal 1 or the drive recorder 3, it performs real-time calculation of the personality score and has a replacement vehicle provision control unit 25 that controls the provision of a replacement vehicle based on the calculated personality score.

To provide a road that allows a vehicle that should be entering the correct lane to enter the wrong lane to return to the correct lane smoothly and safely, reduces congestion, delays, and resulting traffic accidents for vehicles traveling on both the correct and wrong lanes, and maintains a high traffic flow rate. [Solution] The road structure for guiding vehicles traveling in the wrong direction according to the present invention has a first double-lane road (10), a second double-lane road (20), and a central median (30), A predetermined central bypass road (40) is provided in the aforementioned central median (30). It is characterized by having a fence (50) to guide vehicles traveling in the wrong direction and / or a fence (60) to guide vehicles traveling in the right direction.

The invention discloses an unbalanced discrete data generation method fusing causal constraints and potential variables, and the method comprises the steps: firstly collecting environmental factors and accident severity of a road traffic accident occurrence point, recording the environmental factors and accident severity as road traffic accident data D, building a causal graph G of the road traffic accident data D, and obtaining an adjacent matrix A; secondly, performing encoder mapping calculation on the original data feature X through a variational auto-encoder VAE to obtain a potential variable z; taking the potential variable z and the category label Y as the input of a WGAN-GP network generator to obtain a virtual potential variable, and distinguishing the potential variable z and the virtual potential variable in combination with the authenticity score Di output by a discriminator to obtain a high-fidelity potential variable; and finally, taking the high-fidelity potential variable as the input of a VAE decoder, and generating few-class sample synthesis data in combination with causal constraints. According to the invention, the urban traffic accident data monitoring quality can be improved, and the road safety management and decision-making accuracy can be improved.

The invention discloses a 4D traffic accident scene reconstruction method based on an automobile data recorder video, and belongs to the technical field of intelligent traffic. According to the technical scheme of the invention, the method comprises the steps: inputting an automobile data recorder video, extracting video frame features through a pre-trained geometric basis large model DUSt3R, carrying out the frame-by-frame reconstruction and optimization of a 4D scene containing a static background and a dynamic object through the combination of a 3D Gaussian reconstruction technology and dynamic object motion modeling, and finally generating an accident reconstruction video which can be observed at any visual angle. The method solves the problems that the reconstruction scene generalization ability is weak and the static background appearance cannot change along with time in the prior art, realizes the high-precision and high-reality reconstruction of the traffic accident process, and is suitable for the fields of traffic management, accident liability affirmation and the like.