CEO-MLCPP: Collision-Efficient and Obstacle-Aware Multi-Layer Coverage Path Planner for 3D Reconstruction with UAVs

Abstract

In this paper, we propose a novel model-based coverage path planner for the 3D reconstruction of a target structure with an unmanned aerial vehicle(UAV). The proposed method rapidly calculates initial viewpoints considering the ground sampling distance (GSD) by partitioning a structure by height. Then, optimal viewpoints are selected by checking the collision and calculating over-laps and coverage. Next, the newly developed collision-aware Traveling Sales-man Problem (CTSP) is used to connect the optimal viewpoints while guaran-teeing the shortest distance and obstacle avoidance. Finally, the resulting path is refined as a control-efficient trajectory that considers the dynamics of UAVs.The performance of the proposed algorithm is verified by experiments on diverse structures.