Semantic-Driven Informed Planning and 3D Reconstruction for the Quadrotor Unmanned Aerial Vehicle

Autonomous aerial reconstruction tackles the problem of deploying the quadrotor unmanned aerial vehicle in an unknown environment for vision-based surface reconstruction. The state-of-the-art methods mainly extend the classic next-best-view (NBV) framework with the semantic information to concentrate on the reconstruction target. However, the greedy decision of the next viewpoint or branch without global consideration leads to the low efficiency. This paper proposes a novel semantic-driven aerial reconstruction planner, which integrates the task relevance into the planning, making the concentrative reconstruction more efficient. Specifically, a new surface-frontier information structure is extended to maintain the rich semantic information for the fine-grained planning. Driven by an attention-based frontier filter, a novel relevance-aware frontier sequence generation method is proposed to achieve the concentrative reconstruction of the target. Then a semantic-informed local refinement, which incorporates the semantic gain into the graph search, is proposed to further reduce the flight cost and improve the reconstruction quality. Comparative experiments are conducted in simulation environments, demonstrating that the proposed method outperforms the state-of-the-art methods in terms of both the reconstruction efficiency and quality.

I. Introduction

Aerial robots, especially unmanned aerial vehicles (UAVs), are widely deployed in autonomous exploration [1], [2], [3], structural inspection and reconstruction tasks [4], due to their high mobility and autonomy [5], [6], [7]. In a prior unknown environment, the autonomous exploration and reconstruction task requires UAV to iteratively plan the trajectory, obtain the environmental information and build the map based on the onboard sensors and computer. Specifically, exploration focuses on the complete coverage of the environment (maze, cave, etc.) and improving efficiency. Reconstruction focuses on recovering the surface of the buildings and structures, and improving both efficiency and quality. Recent research has been devoted to the UAV planning algorithm for the autonomous reconstruction. However, the issue still remains that the robots can waste time in reconstructing the task-irrelevant parts, especially in complex urban environments. Thus, this paper investigates the leveraging of semantic information for target-focusing enhancement and proposes a semantic-driven informed planner for 3D reconstruction.