A fast obstacle detection method based on optical flow

Abstract

This paper presents a methodology, based on the estimation of the optical flow, to detect static obstacles during the motion of a mobile robot. The algorithm is based on a correlation scheme. At any time, we estimate the position of the focus of expansion and stabilize it by using the Kalman filter. We use the knowledge of the focus position of the flow field computed in the previous time to reduce the search space of corresponding patches and to predict the flow field in the successive one. Because of its intrinsic recursive aspect, the method can be seen as an on-off reflex which detects obstacles lying on the ground during the path of a mobile platform. No calibration procedure is required. The key aspect of the method is that we compute the optical flow only on one row of the image, that is relative to the ground plane.