Detection, tracking and classification of moving objects for unstable video surveillance applications

Detecting, tracking, and classifying moving objects using a single fixed camera is quickly becoming commonplace in many applications ranging from advanced traffic monitoring systems to security applications including border monitoring. In this paper, we present a framework for the detection, tracking, and classification of moving objects in an unstable video taken of a traffic intersection in India. Combinations of methods have been implemented in an attempt to successfully perform the aforementioned tasks automatically. An initial evaluation of these methods is discussed and compared to preliminary results from a manually segmented video.