Measurement of road traffic using spatial frequency analysis of TV images

This paper deals with distinguishing vehicle types in automobile traffic flow by image processing. In the traffic flow, traffic jams always occur. Traffic Jams increase driver stress and result in time delays, traffic accidents, and air pollution. Thus, it is important to pro-vide traffic information such as vehicle volume and to distinguish vehicle types. In this paper, we attempt to do so by using vehicle width. The general measurement of traffic flow is performed using an ultrasonic sensor. But the system is expensive, and the method is not useful for automobile traffic flow. Thus, vue have developed a simple system using a portable camera, a personal computer, and a digital signal processor. This system performs width measurement of vehicles in real time. The measurement method uses the Fourier transform and has the advantage of robustness to noise. Since frequency analysis for width measurement of vehicle images differs from road image analysis, we measure the vehicle volume and distinguish vehicle types in real time. (C) 1998 Scripta Technica, Inc.