A Non-Visual Strategy Study on Boosting Vehicle Surveillance Efficiency for Metropolitan Traffic Network

Metropolitan traffic network is under rapid development in China. Few studies have paid attention to the efficiency of vehicle surveillance system, which is actually at a relatively low level. In this paper, we focus on optimizing the surveillance efficiency through a non-visual strategy, which leads to a much lower computational burden compared to other video-based methods. Our strategy can be considered as a multi-parameter dynamic decision system which includes the road intersection geometry characteristics, traffic lights and traffic flow. A prototype simulation platform is built for massive data study. Simulation experiments demonstrate that our proposed method can effectively boost the vehicle surveillance efficiency in large scale traffic network. Thus reveals the prosperous application value in real metropolitan traffic network.