Sampling-based view planning for 3D visual coverage task with Unmanned Aerial Vehicle

The view planning problem is the problem that involves finding suitable viewpoints for vision-related tasks such as inspection or reconstruction. In this paper, we propose a novel view planning algorithm for a camera-equipped Unmanned Aerial Vehicle (UAV) acquiring visual geometric information of target objects in its surrounding environment. The proposed model-based approach makes use of iterative random sampling and a probabilistic potential-field method to generate candidate viewpoints in a non-deterministic manner. Combinatorial optimization is then applied to select the most suitable subset of these candidate viewpoints to complete the given visual inspection or shape reconstruction task. The effectiveness of the proposed method is demonstrated through a number of computational tests that compare its overall performance against two previous methods. A field-test is also performed to demonstrate the method's applicability in a real world UAV-based shape reconstruction task of an outdoor statue.