Visual-Based Obstacle Detection A purposive approach using the normal flow

In this paper we present a reeex-type behaviour for visual obstacle detection for mobile robots. The main assumption that is made, is that the robot is moving on a ground plane. For the speciic purpose of obstacle detection, it is then possible to design a simpliied vision system that can successfully solve the problem in hand. Under this purposive approach, there is no need to perform any reconstruction of the environment and the system shows a robust performance and is running on simple hardware. The method uses the normal ow information as the input. There is no eeort made in the sense of recovering both components of the optical ow.