Object tracking using particle filter based on Lyapunov stability

Object tracking in video data is a topic that attracts many researchers, many algorithms have been developed in this context. Particle filter is one of the methods that having great success for its characteristic to track object in case of not-Gaussian and non linear system. In this paper we will present a new approach based on the Lyapunov function using the linear matrix inequality formulation. First a motion model is constructed to set the system model of the estimator for estimate the global linear motion. The random work RW model is used to represent the dynamic system and the system's energy is evaluated by the Lyapunov function using the Linear Matrix Inequality (LMI) formulation to establish the estimator. Second we use particle filter to handle the non linear local motion. The dominant color of the moving object in RGB color space will be used as feature to model the appearance of the target. Experiments were performed to confirm the effectiveness of this method to track a moving object.