Image Mosaics Algorithm Based on PSO and SIFT

Abstract

A stable image mosaics algorithm based on PSO and SIFT is proposed in this paper. It adopts SIFT operator to extract corner point feature and finishes gray cross correlation matching, most false matches are eliminated by transcendental knowledge and constraint, then the transform matrix initial value between matches in two images is calculated, the optimal value of this matrix is acquired by PSO, finally, the image mosaics with no gap is accomplished. The experiment results prove the validity and practicability of the proposed method.