Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : May 27, 2020 - May 30, 2020
This paper presents a rapid mapping method for outdoor mobile robot localization using aerial photos taken by a drone. The map is generated as a 3D point cloud constructed by SfM and MVS using multiple pictures taken by the drone camera. An autonomous vehicle equipped with 3D-LiDAR localizes its position and orientation using the NDT (Normal Distribution Transform). Outdoor experiments revealed that NDT works with matching the 3D point cloud from different sensors; camera and LiDAR. Accuracy and precision were compared by different altitudes of the drone where the pictures were taken.
