The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
Online ISSN : 2424-3124
2020
Session ID : 2P1-K03
Conference information

Detection of Traversable Regions in a 3D Point Cloud Map for Accessibility Map Generation
*Sho TATSUGUCHIJun MIURA
Author information
CONFERENCE PROCEEDINGS RESTRICTED ACCESS

Details
Abstract

In recent years, there is a strong demand for increasing the accessibility in public spaces. An accessibility map is a map showing areas where wheelchair users can pass, and it is useful for the disabled and the elderly. In this paper, in order to generate an accessibility map, we estimate traversable regions in a 3D point cloud map. First, a 3D point cloud map is divided into small cells and horizontal planes are detected with consideration of their height. Then, based on the height of each plane, we estimate the connected traversable regions. If there are adjacent cells with a similar height, we define that it is possible to pass between them. In addition, a region where the height is continuously changing is defined as a slope candidate. Finally, we represent their connection relation as an undirected graph.

Content from these authors
© 2020 The Japan Society of Mechanical Engineers
Previous article Next article
feedback
Top