Real-time Depth Estimation for Visual Serving with Eye-in-Hand Robot

아이인핸드로봇의 영상 추적을 위한 실시간 거리측정

Depth between the robot and the target is an essential information in the robot control. However, in case of eye-in-hand robot with one camera, it is not easy to get an accurate depth information in real-time. In this paper, the techniques of depth-from-motion and depth-from-focus are combined to accomplish the real-time requirement. Integration of the two approaches are accomplished by appropriate use of confidence factors which are evaluated by fuzzy rules. Also a fuzzy logic based calibration technique is proposed.