OBSTACLE DETECTION SYSTEM FOR UNDERGROUND MINING VEHICLES

This paper describes an obstacle detection device for autonomous vehic les, such as LHDs and trucks, navigating in mine drifts at speeds of a bout 15 km/h. The system proposed in this paper uses a structured-ligh t range finder and is based upon the direct extraction from images of features related to the obstacle rather than upon a 3D reconstruction of the scene's geometry. The system exploits the temporal evolution of image profiles in order to detect the presence of obstacles and lakes into account phenomena such as ground and wall irregularities, pertur bations of the vehicle's trajectory and impaired visibility conditions .