This paper proposes a visual servo control algorithm with occlusion handling. The proposed control algorithm selects a set of correctly extracted image features, and it estimates all image features from the correctly extracted image features. The estimation procedure enables us to track image features even when occlusion occurs. In addition, even when the algorithm fails to track a moving image feature, it does not need any global search over the entire image plane to find the failed image feature. The image feature selection and estimation are combined with visual servo control in the proposed algorithm. The control algorithm has low computational complexity, since image Jacobian is used for image feature selection and estimation. An experimental result shows the validity of the proposed visual servo control law.
