Research And Implementation Of Vehicle Trajectory Recovery System Based On Traffic Surveillance Camera Data

With the development of Internet of Things technology and big data technology,urban road network has been widely covered by various traffic sensing systems.The vehicle’s trajectories can be accurately sensed by different traffic sensing devices and stored in the data platform.However,the original trajectory data obtained by the traffic sensing system cannot meet the quality requirements of downstream applications.Therefore,this paper proposes a trajectory recovery algorithm based on traffic surveillance camera data to obtain fine-grained,fully-road covered,and fullyindividual-penetrative trajectories and implements a complete trajectory recovery system.In the design of vehicle trajectory recovery method,this paper first performs trajectory preprocessing based on the urban road network,through three steps of urban road network processing,traffic surveillance camera trajectory generation,and GPS trajectory road network matching.Then,this paper designs a vehicle trajectory recovery algorithm,which consists of four effective modules.By calculating the similarity between GPS fine-grained trajectory and traffic surveillance camera coarse-grained trajectory movement patterns,modeling the turning probability information of similar fine-grained trajectories at intersections,extracting movement pattern information in coarse-grained trajectories,and finally combining the two types of information to achieve vehicle trajectory recovery.The effectiveness of the proposed method is demonstrated through experiments based on the real dataset of a Chinese city.In the implementation of the vehicle trajectory recovery system,this paper designs five effective modules to complete data access,real-time computation,offline computation,backend querying,and frontend visualization display.On the other hand,in order to improve system performance,this paper separately designs real-time computation processes and offline computation processes to accomplish tasks with different functional requirements.In this system,vehicle trajectories generated in real-time in the city can be effectively sensed and recovered,providing city managers with complete vehicle movement information and supporting numerous downstream applications effectively.