Abstract:

An crucial problem of feature-based image registration is how to get enough and accurate correspondent features to increase the accuracy of registration. An automatic remote sensing image registration method based on optimized energy analysis of control points/correspondences is presented to improve the traditional algorithm.First,a multi-scale segment method based on digital curvelet transform is used to evaluate the quality of sensed images, then parameters of SIFT feature-matching algorithm is automatically adjusted according to multi-scale grey relation energy of image quality  to increase the feature matching points and the correspondences.At last, to make sure the accuracy of geometrical transform parameters, a method of selecting optimized correspondences based on the distribution and matching energy model of control points is developed. Experimental results demonstrate that the proposed method works well in increasing control points and the correspondences of low quality remote sensing images.

Key words: Remote sensing, Image reGistration, Correspondences selection, Scale invariant feature transform (SIFT), Grey relation analysis