Image Mosaic Algorithm Based on Vertical Distance of Edges

This paper presents a new algorithm of automatic image mosaic, which generate mosaic image by comparing the vertical distance of edges in two neighboring images. This algorithm selects points to be matched by statistical features of the images before the actual matching stage, reducing the amount of computation, saving time for higher accuracy. Based on vertical distance of edges, the matching stage of this algorithm is able to locate the optimal joint, which bears high consistency in two neighboring images. To stitch two images on the optimal joint generates a stitching line that is less noticeable. This feature makes it unnecessary to smooth large area to merge two images, leaving more untouched details in the mosaic image. The algorism also computes the vertical displacement of the two images, so it is able to deal with images shot by unfixed cameras.