Measurement of Vehicle Queue Length Based on Video Processing in Intelligent Traffic Signal Control System

The research of intelligent traffic signal control system has recently received much more attention due to the increase of traffic jams and accidents. In case of the controlling parameters, traffic flow counted in the green cycle is widely adopted by researchers to adjust the next green period. Those algorithms can´t achieve real-time control for they always have one period delay. In this paper, authors make use of vehicle queue length during red cycle to control the green cycle in the same period. A method of detecting the length is proposed by taking the texture difference to separate vehicles from roadway as well as edge detection. Experimental results indicate that the approach is simple and acts stably, maybe the foreground of application is broad.