Estimating Traffic Flow by Image Processing and the Usage of an Adaptive Traffic Signal Control System

The Robust and reliable traffic information detection from images acquired by traffic surveillance system is an important problem with numerous applications including adaptive traffic signal control system and traffic analysis system. This paper describes an application of computer vision techniques to traffic surveillance. The objective in this paper is to improve the efficiency of traffic flow estimation that include pedestrian flow and vehicular flow estimation. Specifically, some vision based methodologies that including perspective transformation, object segmentation and classification have been employed to traffic flow estimation. Traffic congestion in urban area is still an important issue to be solved. One of the reasons of congestion in urban area is imperfect operation of traffic signal control. A traffic signal control problem can be formulated as a decision making problem for a stochastic dynamic system. To minimize the queue length and vehicle delay time, an adaptive traffic signal control system that used signification traffic information to find the optimal phase sequencing of traffic signal has been proposed in this paper. The experiment results of intersection traffic simulation indicate that our approach has better performance than other two control systems.