Research on Image Feature Matching Algorithm Applied to UAV Obstacle Avoidance

摘要

The ORB, SIFT, and SURF algorithms are currently common feature matching algorithms. These three algorithms have certain differences in execution speed and matching robustness. In this paper, the data sets provided by Mikolajczyk and Schmid are used to compare the algorithms in terms of execution speed, image transformation robustness, and noise robustness. The conclusion of this paper is that the ORB algorithm is the fastest, the SURF algorithm is the most robust, and the SIFT algorithm has the most feature points.