Abstract
Recent years, the number of potential wheelchair users is acceleratingly growing in Japan, as facing the issue of rapid aging society. Barrier-free maps, for example, are prepared in many municipalities, but they miss the details of physical barriers, such as bumps and width clearances on the path, needed for the independent mobility of the wheelchair users. This paper addresses a method to find out physical barriers by using a depth camera, so the facility managers easily check and improve the environment by themselves. The proposed system robustly estimates the floor surface from the point cloud captured by a depth camera and RANSAC algorithm and highlights the existence of interference between the environment and the volume of an actual wheelchair would occupy to pass through. This paper demonstrates the achieved accuracy of the barrier checks in an actual public space.
