Accurate Autonomous UAV Landing Using Vision-Based Detection of ArUco-Marker

This paper presents a solution of autonomous accurate landing problem for an unmanned aerial vehicle (UAV) in the target point. The combination of two algorithms for on-image search and accurate landing on marker is presented. The UAV landing on marker includes the unit for adjustment of aerial vehicle position relative to the marker to increase the landing accuracy. Also, the paper shows the on-image marker detection algorithm, to define position and alignment of UAV camera relative to the marker. The developed algorithms were tested in the Gazebo simulation environment in real-world settings. According to the experiments, these algorithms outperform in accuracy than similar systems for automated landing, based on computer vision (5–10 cm) and GPS-based UAV control systems (2–5 m). Landing error of UAV in simulator with adjustment was 19.75 mm, in real-world settings – 21.2 mm.