Image registration method based on SURF and FREAK

This paper proposes a new method for image registration by combining SURF and FREAK. SURF can extract robust feature points, and the topology of FREAK descriptor has strong ability of regional description. First, feature points of images are extracted by SURF, and described by FREAK descriptor. Then descriptors are roughly matched through the ratio of the closest neighbor and second closest one. Second, use RANSAC to eliminate the wrong matches. At last, the best transform parameters of spatial geometric are obtained by least squares method. Experimental results show that this method has a good matching result, and the speed of operation has been further improved under the premise of ensuring registration accuracy.