Image Mosaicing Using Homography Estimation

In this article, we propose to join multiple images using the homography estimation technique. First, the images are classified, then the pairwise homography between the images is calculated, and then a combination of images is performed with the lens. Then a new decision fusion framework based on a spanning tree and Exif data is proposed to obtain the order of the images. After obtaining the order of the images, an estimation of the projective homographies is made in pairs between the images. Then we deform the images according to the computed homography. Deformed images are later faded and laplacian mixed for seamless seams.