Traffic data collection is crucial for the management of road networks. Based on YOLOv5s and DeepSORT, optimizations are made in the vehicle detection stage and the vehicle tracking stage to improve the accuracy of the system. At the same time, Support Vector Machine(SVM) is introduced to divide the upper and lower lanes, and the reference lines are automatically defined for videos taken from different perspectives. This solves the shortcomings of traditional vehicle counting methods that need to use manual annotation for each video to distinguish upper and lower lanes and set reference lines. Finally, the experimental results show that the system is able to achieve an average accuracy of 95.5% in real time in daytime, evening, foggy, and sunny conditions.
A Vehicle Counting and Road Condition Analysis System Based on Multiple Object Tracking
Lect. Notes Electrical Eng.
Chinese Intelligent Systems Conference ; 2022 October 15, 2022 - October 16, 2022
27.09.2022
8 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Road condition detection system based on vehicle trajectory tracking
Europäisches Patentamt | 2023
|Object Detection and Tracking Algorithms for Vehicle Counting: A Comparative Analysis
ArXiv | 2020
|Vehicle Counting System using Deep Learning and Multi-Object Tracking Methods
Transportation Research Record | 2020
|Europäisches Patentamt | 2023
|Europäisches Patentamt | 2023
|