A single-camera feature-based vision system for helicopter autonomous landing

In this paper a feature based single camera vision system for the safe landing of an unmanned aerial vehicle (UAV) is proposed. The autonomous helicopter used for tests is required to navigate from an initial to a final position in a partially known environment, to locate a landing area and to land on it. The algorithm proposed for the detection of safe landing areas is based on the analysis of optical flow and of mutual geometric position of different kinds of features, observed from different points of view. Vision allows estimating the position and velocity of a set of features with respect to the helicopter while the onboard, hierarchical, behavior-based control system autonomously guides the helicopter. Results, obtained using real data and a real helicopter in a outdoor scenario, show the appropriateness of the vision-based approach. It does not require any artificial landmark (e.g., helipad), is able to estimate correctly and autonomously safe landing areas and is quite robust to occlusions.