Design and Implementation of an Intelligent Traffic Management System Based on Improved YOLO and Multi-Modal Fusion
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Aiming at the problems of insufficient target detection accuracy, poor multi-target tracking stability, and rigid traffic signal control in complex environments of urban traffic scenarios, this paper proposes an intelligent traffic management system that integrates improved YOLO series algorithms and multi-modal perception technology. With the DB-YOLOv12 model as the core detector, the system combines the improved DeepSORT tracking algorithm to achieve high-precision recognition and continuous tracking of targets such as vehicles and pedestrians. It also integrates a multi-modal weather recognition module and a dynamic traffic signal control strategy to construct a closed-loop management mechanism of "perception-analysis-decision". Experimental results show that the system achieves a vehicle recognition accuracy of 94.7% in complex environments, a pedestrian detection recall rate of 82.1%, and a foggy weather recognition accuracy of 87.5%, which can improve the traffic efficiency of intersection during peak hours by more than 30%. The research results provide important technical support for the engineering application of intelligent transportation systems.
Fields of Study & Subjects
Advanced Neural Network Applications
Video Surveillance and Tracking Methods
Autonomous Vehicle Technology and Safety
Publication details
- Journal / source
- IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
- Pages
- 1149–1154
- Publication date
- 28 January 2026
- Language
- EN