Real-time obstacle detection with a single camera

In this paper, a region-based obstacle detection algorithm for indoor navigation is presented. Firstly, the image is segmented into regions. Then each individual region is classified as belonging either to an obstacle or the ground based on its average 'disparity'. The proposed method only uses a single passive color camera, and can provides a local obstacle map at high resolution in real-time. Experimental results are given for real indoor scene images to validate the efficiency of this method.