Remote sensing images registration with different viewpoints

Abstract

Image registration techniques in remote sensing (RS) images play an important role in the field of geographic information system. However, the registration accuracies using current techniques are significantly affected when ground relief variations and imaging viewpoint changes. In this paper, we propose an image registration method for RS images with different viewpoints. We represent two RS images taken from different viewpoints (or affected by ground relief variations) using their extracted SIFT feature points, and register them by the proposed mixed feature based non-rigid correspondence estimator for solving the problems of ground relief variations and imaging viewpoint changes. The image transformation model for finally mapping two images is defined by an affine transformation. We evaluated the performances of the proposed method by two sets of RS images obtained from the Google earth and the unmanned aerial vehicle (UAV), and compared with three state-of-the-art methods where our method shows the best alignments in most cases.