A knowledge-based objects tracking algorithm in color video using Kalman filter approach

A new knowledge-based algorithm for the purpose of rigid and non-rigid objects tracking through color feature in video sequences is proposed in this research. The mean shift (MS) algorithm, as the efficient method in the area of color-based objects tracking, is improved to solve the tracking problems, such as background with similar colors, partial or full occlusion, sensibly, and so on. In the algorithm presented here, an improved convex kernel function is realized to present a particular solution, since a robust estimator, i.e., Kalman filter approach is correspondingly realized. Experimental results verify that the proposed algorithm is efficient, under sever conditions, while the speed of the object could be constant.