An automatic video image mosaic algorithm based on SIFT feature matching

Abstract

An automatic video image matching algorithm based on SIFT feature points was presented in this paper. The optimized implementation of the best-bin-first(BBF) algorithm based on kd-Tree was used in coarse matching, random sample consensus(RANSAC)was used in accurate matching, the last use the global information are calculated to update the match precision. with pixel brightness weighting in HIS color space was proposed in blending images. The experimental results show that the method with strong robustness performs fast and effectively and has highly valuable in practice. © 2013 Springer-Verlag.