Improvement in extended object tracking with the vision-based algorithm

Abstract

The main emphases on any object based tracking with vision algorithms are parametric state space based algorithms like a Bayesian filter and its family of algorithms or nonparametric algorithms like Mean Shift algorithms which are color sensitive. In this paper, We have considered vision based algorithm with Bayesian filter algorithms. We have seen more of the state space tracking algorithms uses point based object tracking approaches in which researchers did very well in the last decades, with the advent of faster computing devices available the tracking algorithm have improved a lot with extended object tracking where the object is tracked as an entire object instead of the point-based approaches. The proposed system of the algorithm is providing good results in terms of state estimation over its point based approach. So using this vision algorithm the complete object can be tracked using sensor system and this is the novelty of paper.