In this paper, we propose the optimal estimation of motion parameters and structures ofa three-dimensional(3-D) rigid object from a monocular image sequence. First, object motion model and camera imaging model are discussed. And then, through further analysis, the state equation and measurement equation of the object is established, an iterated extended Kalman filtering(IEKF) algorithm is adopted when the estimated system is continuous and nonlinear. The simulative numerical experiments have shown that the presented method is efficient. Keywords: 3-D object recognition, motion parameters and structures estimation, monocular image sequence, motion model, imaging model, Kalman filtering
|