Video Based Vehicle Tailgate Behaviour Detection in Urban Road Junction

This paper proposes an new algrithem framework for video based vehicle ‘tailgate' behavior detection in urban road junction. Based on the road traffic regualtion about the illegal parking behavior definition, the author proposes traffic lights signal monitor, vehicle tracking and road conjestion detection algrithems in the real time video analyse. The measured parameters including vehicles' trojactory, speed, and traffic status are used for the traffic violation behaviour identification. The algrithem and system are applied at the road junction of Huashan/Taian road, Shanghai, China, which achieved 90% accuracy in traffic violance detection.