Satellite image registration using hybrid salient region detection method

摘要

In remote sensing multisensor image fusion or registration is the process of combining relevant information from two or more images into a single image. The resulting image will be more informative than any of the input images. In order to transform the remote sensing images to retrieve more information, this paper proposes a hybrid method that consists of image segmentation, salient region detection and image fusion. First of all, the paper presents the superpixel segmentation method in order to divide the image into subareas and for the feature extraction we implemented the difference of Gaussian and local binary pattern from the salient regions. This proposed method is tested on remote sensing images. Software results shows that the method is fast and gives less error compared to other existing methods.