Man machine cooperation for 3D objects pose estimation

Geometrical modelling methods are of great interest to enhance the control of telerobots. In this article we develop a method based on a man-machine cooperation to position 3D volumes over their 2D images issued by on-board cameras. The human operator brings the object in rough correspondence with a selected image detail. Then fine positioning is automatically obtained. The autonomous refinement method is described and results are discussed.<
>