Wavelet Multi-Resolution Image Mosaic Combines SIFT Feature Points and Region Features
Chaoliang Zhang, Hanhong Jiang, Chunliang Jiang, Yang Wu, Chongyuan Hou, Jianxuan Li · 2009
This paper researched scale invariant feature transform (SIFT) which is feature extraction method of image processing, and it is stable for image feature registration. And image mosaic needs the fusion of overlapped region after the image registration. The research of image fusion in mosaic nowadays is concentrated on multi-resolution mosaic simply, and ignores the similarity of feature points among the pictures, with part of specific information lost. So this paper developed a wavelet multi-resolution image mosaic method combined with SIFT feature points and region features. The experimental results show that this method had an advance in fusion performance evaluation and improved the quality of image mosaic.