The object tracking algorithm is used to tracking multiple objects in a video streams. This paper provides
Mutual tracking algorithm which improve the estimation inaccuracy and the robustness of clutter
environment when it uses Kalman Filter. using this algorithm to avoid the problem of id switch in
continuing occlusions. First the algorithms apply the collision avoidance model to separate the nearby
trajectories. Suppose occurring inter occlusion the aggregate model splits into several parts and use only
visible parts perform tracking. The algorithm reinitializes the particles when the tracker is fully occluded.
The experimental results using unmanned level crossing (LC) exhibit the feasibility of our proposal. In
addition, comparison with Kalman filter trackers has also been performed.