ABSTRACT
Robot vision can be interpreted as that part of computer vision that is ultimately intended to be used to guide the actions of robots. In the context of intelligent robotics, robot vision can be thought of as a powerful perceptual component. The sensory instrumentation may be fixed in the environment or mounted on the robot, whether on its body or hand. The on-board rangefinder gathers position data from all surfaces visible all around the robot up to some maximum range measurable by the rangefinder. The continuity of the robot’s movement can simplify this process by constraining the search space of the match task. A more elegant approach would be to identify appropriate landmarks and to recognize these same landmarks as the robot moves; this clearly requires more of the vision system. Many robot vision ranging schemes have been inspired by human range extraction mechanisms.
