Aiming to address the problem of detecting traffic flow in the various complex traffic environments, this paper proposes detection and tracking methods of multi-target moving vehicles based on CenterNet and CenterTrack respectively, and evaluates its performance with public KITTI dataset and four self-collected datasets. The experimental results show that moving vehicles can be effectively detected and tracked in real time under different traffic environments including nighttime, rainy and crowded scenes, the proposed method can reach 97%average detection accuracy with nearly 26fps of tracking speed and is capable of dealing with different traffic and climate conditions. Compared to previous methods, the method proposed in this paper can detect and track the traffic flow more quickly and preciously.
Real-time detecting of urban traffic flow based on deep learning
5th International Conference on Computer Information Science and Application Technology (CISAT 2022) ; 2022 ; Chongqing,China
Proc. SPIE ; 12451
20.10.2022
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Urban traffic flow prediction method based on deep learning
Europäisches Patentamt | 2025
|Deep Learning Mechanism for Region Based Urban Traffic Flow Forecasting
Springer Verlag | 2023
|Urban road traffic flow prediction method and device based on space-time deep learning model
Europäisches Patentamt | 2020
|