Tracking the Human Motion from Multiple Video Images

Abstract

Object motion and human motion tracking is an important technique used in many systems, especially in the fields of computer vision. An automatic motion tracking and catching system is developed in this research. It can be used to entrap snatchers for shopping mall, office, housing, corridors, borders, secured buildings, at airport and so on. The multiple perspective imagery is applied to track the desired people successfully. It needs to establish the perspective Gaussian Mixture Model and Particle Filter calculation correspondence among the multiple cameras. At once a man pass and is marked or chosen to track any view of camera which is implemented, the system will track the motion of this subject and as well as the other cameras will track him. In this paper, we can track and recognize the subject we desire.