Robust invariant automatic image mosaicing and super resolution for UAV mapping

Abstract

The aim of this paper is to develop efficient, robust and automated image frame fusion algorithms for mosaicing and super-resolution. Image registration is the key step in combining multiple independent low-resolution images to give one large mosaic image with high resolution. Our research is based on Scale Invariant Features Transform (SIFT detector/descriptor) for geometric and photometric images registration for mosaicing. The proposed super-resolution technique is also based on SIFT features and integrated to the image mosaicing in order to construct a full high resolution UAV map that is insensitive to the order, orientation, scale and illumination of the images. Real data validations show the effectiveness of our proposed techniques in comparison with other classical techniques.