Semantic SLAM-based Autonomous Tributary Navigation System Using 3D LiDAR Point Cloud for UAV

Remote sensing methodologies are useful for managing natural systems. Among them, we focused on a methodology for collecting data by attaching three-dimensional (3D) light detection and ranging (LiDAR) sensors to unmanned aerial vehicles (UAVs) that can perform detection over a wide area with high mobility. In this study, we propose a semantic simultaneous localization and mapping (SLAM)-based autonomous driving system for tributary mapping by using a UAV equipped with a 3D LiDAR. As a natural system, tributaries are not studied extensively for diagnosing polluted watersheds and managing water quality. Therefore, it is necessary to develop a robust autonomous driving system for mapping tributary environments.