Abstract:

Based on the advantage of non-subsampling contourlet transform (NSCT) in image decomposition and improved SIFT algorithm,a new method of registering remote sensing images is proposed in this paper.First,the remote sensing image to be matched and the reference image are decomposed by NSCT.Two low-frequency images of remote sensing images are acquired.Secondly,the Canny edge detection is used to require the edge coordinates of image,this points is contrasted with the coordinates of the key points detect by SIFT which can remove some unstable response points.Finally,use the RANSAC algorithm to remove the wrong point,the rest of the points are used to match remote sensing images.The experimental Results show that the new method can efficiently decrease remote sensing image matching time by only use SIFT algorithm and increase the accuracy of remote sensing image matching.

Key words: remote sensing image, improved SIFT algorithm, NSCT, image matching