Novel Image Fusion Technique Based On DWT & MSVD

Abstract

Image fusion is the process of combining two or more images with specific objects with more clarity. It is common that in focusing one object remaining objects will be less highlighted. Hence to get an image highlighted in all regions, a different means is required. This is what is done by the Image Fusion. In remote sensing, the increasing availability of Space borne images and synthetic aperture radar images gives a motivation to different kinds of image fusion algorithms. In the literature a number of time domain image fusion techniques where the substitution operations are done pixel value by pixel value. Few transform domain fusion techniques are proposed. In transform domain fusion techniques, the source images will be decomposed, then integrated into a single data and then reconstruct back into the time domain. In this paper, new transform techniques like singular value decomposition will be utilized for image fusion. In the literature, the quality assessment of fusion techniques is mainly by subjective tests. In this paper, objective quality assessment metrics are calculated for existing and proposed techniques. It has been found that the new image fusion technique outperformed the existing ones.