Using space continuity and orientation constraints for range data acquisition

摘要

Fast range data acquisition of objects is often needed in computer vision applications such as robot vision. In this paper, a fast range data acquisition approach based on space continuity and orientation constraints is proposed. First a masking technique which can reconstruct objects efficiently with the epipolar constraint in trinocular vision is presented. Subsequently the artificial point problem is analysed and an artificial point removal algorithm based on a space continuity constraint is proposed. In order to ensure correct initial reconstructed points, an orientation constraint is introduced. Finally, some reconstruction experiments illustrate the feasibility of our approach.