Next-Best View Policy for 3D Reconstruction
This addresses the challenge of efficient 3D reconstruction for drone operators, though it appears incremental by applying reinforcement learning to an existing problem.
The paper tackles the problem of incomplete 3D models in drone-based reconstruction by proposing Scan-RL, a learning-based algorithm for Next-Best View policy, which reduces steps and distance compared to a baseline circular path.
Manually selecting viewpoints or using commonly available flight planners like circular path for large-scale 3D reconstruction using drones often results in incomplete 3D models. Recent works have relied on hand-engineered heuristics such as information gain to select the Next-Best Views. In this work, we present a learning-based algorithm called Scan-RL to learn a Next-Best View (NBV) Policy. To train and evaluate the agent, we created Houses3K, a dataset of 3D house models. Our experiments show that using Scan-RL, the agent can scan houses with fewer number of steps and a shorter distance compared to our baseline circular path. Experimental results also demonstrate that a single NBV policy can be used to scan multiple houses including those that were not seen during training. The link to Scan-RL is available at https://github.com/darylperalta/ScanRL and Houses3K dataset can be found at https://github.com/darylperalta/Houses3K.