Identifying Background Features Using LiDAR
Published in U.S. Patent Application Publication US 2021/0373173 A1, 2021
This patent introduces a method for identifying background structures in LiDAR point cloud data to isolate foreground objects and extract drivable areas. The approach models the 3D environment as a spherical structure and segments point clusters into faces. A graph-based representation of these faces is constructed and analyzed to detect subgraphs representing background surfaces. The technique enables robust background segmentation critical for navigation and terrain understanding in autonomous systems.
Recommended citation: Özaslan, T. (2021). "Identifying Background Features Using LiDAR." U.S. Patent Application Publication US 2021/0373173 A1.
Download Paper
Related people: Tolga Ozaslan