Abstract: The design and detection of landmark are key points in the process of unmanned aerial vehicle (UAV) autonomous landing based on computer vision. A fast contour corner detection algorithm is proposed in this paper and a new nested triangle pattern is also designed as a landmark for autonomous landing of UAV. Firstly, the background and target overall contour information, obtained by Suzuki-Abe algorithm, are used to extract the nested contours. Secondly, the corners are detected by the improved Douglas-Peucker fitting algorithm. Due to the optimization of the corners number and the searching method for the farthest two points, the computational time and the complexity are reduced greatly. The experimental results show that the proposed technique performs well when the distance from the landmark is relatively far and the part of landmark information is missing, and it can meet the demand of the real time measurements of the position for the autonomous landing process of UAV.

Key words: unmanned aerial vehicle (UAV), autonomous landing, contour extraction, corner detection, landmark