An improved algorithm of multi-source remote sensing image registration based on SIFT and Wavelet Transform

Multi-source remote sensing image registration is a challenging task in remote sensing. Considering the fact that multi-source remote sensing image have characteristics of various image properties, low correlations and matching distributed nonuniformity spatially. This paper proposes an improved algorithm based on SIFT[1][2] and Wavelet Transform to ameliorate these problems. By using Wavelet Transform, First the image will be decomposed into low-frequency sub-band and high-frequency sub-band. Next we extract the SIFT feature points of the low-frequency and the image´s edge feature points of the high-frequency by curvature scale space(CSS)[3] corner detection. Finally we choose the fine matching points to realize the image registration using the improved random sample consensus(RANSAC).