Downsampling of a 3D LiDAR Point Cloud by a Tensor Voting Based Method

Abstract

A point cloud is a data format that consists of a combination of multiple points used to identify an object or environment. Point cloud registration is related with many significant and compelling 3D perception problems including simultaneous localization and mapping (SLAM), 3D object reconstruction, dense 3D environment generation, pose estimation, and object tracking. The aim of this study is to ensure that the point clouds obtained with 3D LiDAR are sampled while preserving their geometric features. For this process, it is inspired from the method known in the literature as Tensor Voting, which is used to extract geometric features in N-dimensional space. After determining the areas with high density in the point cloud, with the help of tensor voting, it is aimed to express the same geometric attributes with a lower number of points by reducing the density.