In this thesis, a visual servoing system for a hand held camera robot is investigated. For visual servoing, a new concept. called augmented image space, is adopted, along with the method of the inverse Jacobian. In the method, the task is described in the 3-D augmented image space with the depth information incorporated. System software is developed based on PROLOG and a set of robot motion commands is supplied as a library. It is shown that task descriptions in the image space can be easily programmed by using the proposed software system. And to demonstrate effectiveness of the developed system, experiments were performed.
