The study of the detection and tracking of moving pedestrian using monocular-vision

Abstract

To ensure the safety and efficiency of the pedestrian traffic, this paper presents a real-time system for moving pedestrian detection and tracking in sequences of images of outdoor scenes acquired by a stationary camera. The self-adaptive background subtraction method and the dynamic multi-threshold method were adopted here for background subtraction and image segmentation. During the process of tracking, a new method based on gray model GM(1,1) was proposed to predict the motion of pedestrians. And then a template for tracking pedestrian continuously was presented by fusing several characters of targets. Experimental results of two real urban traffic scenes demonstrate the efficiency of this method, then the application of this method is discussed in real transportation system. © Springer-Verlag Berlin Heidelberg 2006.