An efficient SIFT-based matching algorithm for optical remote sensing images

Scale-invariant feature transform (SIFT) and its improved versions have been widely used to match the remote-sensing optical images. However, it is still a challenging task to get enough correct matching pairs by using SIFT-based methods. In this letter, an efficient matching scheme is proposed to increase the number of correct matches. The proposed matching scheme uses scale, orientation, and translation differences between the input images to remove the outliers and to retain the correct matches. At first, the initial matching candidates are obtained by a SIFT-based algorithm. Then, our proposed matching scheme is used to select the correct matching pairs. The proposed method can obtain more correct matches than the state-of-the-art methods. Experiments are performed on five pairs of optical images to verify the performance of the proposed method.