Urban road intersections are the key nodes of urban road network with a mass of traffic flow confliction, which often result in traffic accidents. Therefore, automated conflict detection is crucial for traffic safety analysis. This paper proposes a method to extract traffic conflicts by using deep learning based trajectory detection. The traffic video data is collected by unmanned aircraft in a road intersection in Xi’an, i.e., an ancient city in China. Then, YOLOv5-DeepSORT is employed to extract the vehicle trajectories. Comparing the situation and velocity of vehicles in each time slice, we put forward a method of automatic calculation and extraction of traffic conflict indicator. Moreover, a visualized conflict areas are shown in graphs, which is helpful for the safety analysis of urban road intersections.
Extracting Traffic Conflict at Urban Intersection Using Deep Learning Trajectory Detection
Lect. Notes Electrical Eng.
International Conference on Autonomous Unmanned Systems ; 2022 ; Xi'an, China September 23, 2022 - September 25, 2022
Proceedings of 2022 International Conference on Autonomous Unmanned Systems (ICAUS 2022) ; Chapter : 282 ; 3057-3069
2023-03-10
13 pages
Article/Chapter (Book)
Electronic Resource
English
Network-Wide Vehicle Trajectory Prediction in Urban Traffic Networks using Deep Learning
Transportation Research Record | 2018
|Distributed Multi-Intersection Traffic Flow Prediction using Deep Learning
DOAJ | 2024
|Using Traffic Conflict Method in Evaluating Traffic Safety at the Reconstructed Intersection
DOAJ | 2002
|Intersection traffic distribution and trajectory calculation method for traffic simulation
European Patent Office | 2024
|