A Traffic State Detection Tool for Freeway Video Surveillance System

Abstract

Traffic state is one of the most important traffic flow parameters to both the traffic management center and the traveler. It's difficult to extract traffic data using surveillance cameras because of the wider field, panning and zooming of the surveillance cameras. To leverage the existing surveillance camera infrastructure, a surveillance video based traffic state detection system is proposed. The proposed system can estimate traffic flow speed and road space occupancy, and recognize three typical traffic states (congested, slow, and smooth). Experimental results show that the system had good adaptation and high accuracy in daytime.