An efficient sequence image automatic mosaic algorithm
Guixi Liu · Journal of Optoelectronics·laser · 2011
In this paper,an efficient sequence image automatic mosaic algorithm based on phase correlation and speeded-up robust features(SURF) algorithm is proposed.Firstly,the peak value and the translation offset of the normalized cross-correlation function between adjacent images were computed and the disordered image sequence was sorted by maximal correlation degree intersecting.Then,the region of interest(ROI) was roughly computed,which improves SURF feature points extraction method and speeds up the matching stage using both fast nearest-neighbors search and random sample consensus(RANSAC) algorithm.Finally,singular value decomposition(SVD) method was used to determine the transform model,and with the frame-to mosaic stitching method and weighted average blending method,the panorama was generated.The experimental results indicate that the proposed algorithm with strong robustness and high value not only exceeds existing ones at speed but also can stitch the disordered image sequence into a panoramic image effectively and automatically.