ABSTRACT

There is a vast amount of research in the field of computer vision, which has great potential in various fields of engineering, in particular robotics and automation. Computer vision units are effectively used in industrial robotics for inspection and handling of manufactured parts. Most of these systems operate in a controlled environment with good illumination and minimum interference. However, mobile robots mostly operate in unpredictable environments. Typically, a vision unit is employed for steering a robot, avoiding obstacles, detecting landmarks, handling objects of interest, tracking objects, or navigation (Chen and Tsai 1997; DeSouza and Kak 2002; Kosaka and Kak 1992; Kumagai and Emura 2000). In early mobile robots, visual feedback was commonly used for robot navigation and obstacle avoidance. Recently, particularly in service robotics, vision is used for interacting with people, recognizing the environment and handling objects. Needless to say, vision is one of the most desirable features that we would like to have in a robot.