3D point cloud model reconstruction method by using multi–view 2D images and 3D point clouds
Yukihiro TODA, Koichi Hashimoto · The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) · 2018
This paper proposes a 3D point cloud model reconstruction method by using multi-view 2D images and 3D point clouds of an object. In order to reconstruct a point cloud of a 3D shape, point clouds registration is required. However, keypoints and correspondences between point clouds could not be detected using a shape with less feature. Therefore, we propose a method using 2D image-matching in order to overcome this problem. 2D image and 3D point cloud taken from the same view point of an object and utilized to find the correspondences between a point in 3D space and a position in the 2D image.