Artificial landmark design and detection using hierarchy information for UAV localization and landing

Abstract

Localization and landing are of fundamental importance for Unmanned Aerial Vehicles (UAVs), and with the research on compute vision information processing, artificial landmark detection, a more efficient and accurate method, becomes a hot topic. In this paper, in order to provide sufficient information for localization, we design an artificial landmark with an annulus and some circles inside with specific positions and colors. We firstly use color and hierarchy information to find possible patter, then calculate its cross-ratio to check this pattern, localize our UAV at last. To prove our landmark and detection-recognition algorithm, two kinds of experiments are conducted and the results demonstrate our feasibility and accuracy. One is putting our landmark on random backgrounds to imitate complex environment, another is on real environments with different projection and distance. With those experimental results, we get the conclusion that our landmark and algorithm are robust and feasibility under various environments.