A Robust and Efficient Airborne Scene Matching Algorithm for UAV Navigation

The purpose of this paper is to introduce a novel airborne scene matching based unmanned aerial vehicle (UAV) navigation technology in case of global navigation satellite systems (GNSS) fail to work. Our primary motivation focuses on reducing the dynamic UAV positioning errors in traditional scene matching algorithms, which are caused by real-time image distortions when UAV maneuvering flight. Firstly, a novel aerial image matching technique based on simplified haar-like local binary pattern (SHLBP) is proposed to obtain the position of matching points. Then random sample consensus (RANSAC) is applied to remove mismatches by iteratively minimizing the average residual. Finally, the UAV position is calculated according to the position of matching points in the UAV motion model and pinhole imaging model. The proposed method is tested on real flight-test data. The experimental results have demonstrated that it can obtain accurate UAV position by the proposed method. Compared with other state-of-the-art aerial image matching algorithms, the proposed algorithm has better performance in matching precision and computational efficiency.