Abstract: The design and detection of landmark and unmanned aerial vehicle (UAV) pose estimation are key points in the process of UAV autonomous landing based on computer vision. This paper proposes a pose estimation algorithm based on polar coordinates transform and designs a new landmark. Firstly, the polar coordinates transform is applied to the pose estimation algorithm for its related parameters are few and low computational complexity, then a landmark with hemi-toroidal shape is designed. Secondly, the saliency detection algorithm is used to detect the landmark, which is accurate, fast and robust to clutter background. Then the Hough transform algorithm is used to extract the lines of the detected landmark. Lastly, the experimental results show that the landmark is feasible in the process of UAV autonomous landing.