A tracking technology of feature point's fuzzy clustering algorithm based on image stabilization
Dandan Qu, Yuejin Zhao, Bin Wang, Mei Hui · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2008
Object tracking technology combined with image stabilization is called tracking technology based on image stabilization. Moving objects affect stabilization compensation in tracking algorithm based on image stabilization. The middle value method, former background method and dynamic clustering method are not useful for rotate or non-rigid objects. Fuzzy Clustering of feature points is proposed to solve these problems. First background and objectives membership value of the pixel are calculated, and then pixels are classified to background and target categories accurately by defining membership threshold of the background and targets. Experimental results show that fuzzy clustering algorithm resolves the moving target interference problem in image sequence and realizes steady tracking of object. It is proved to be more robust and less sensitive to the numbers of initial clustering.