A video vehicle tracking algorithm based on ploar Fourier descriptor

Abstract

Video-based vehicle tracking movement in the field of intelligent transportation is an important research question. This paper presents a video vehicle tracking algorithm based on polar Fourier descriptor according to the Fourier transform properties. This paper first gets the ideal vehicle for the target object and extract its contour by using background subtraction for the reason that the traffic monitoring video scenes are fixed and then gets the video frequency spectrum characteristics of the vehicle by implementing Discrete Fourier Transform to the image data in polar coordinates and then achieves the video vehicle tracking based on linear prediction and shape similarity measurement. Experimental results show that this algorithm's tracking accuracy is higher and time-consuming is less. © 2011 Springer-Verlag.