An Improved Algorithm for Image Mosaic

This paper puts forward a novel method using dynamic programming and grey relational analysis to find a best seam line because of the grey scale difference in the edges of image in the overlapping region and the relation of neighboring points. The first step is to find an optimal seam line to stitch the images. Then the seam line is removed by a hard correction method based on dynamic width. Finally, the seamless mosaicking image is created. Three-kind different experimental results of regular texture images, irregular texture images and different temporal images indicate that the new method is very effective, simple, easy to realize and can gain more satisfying mosaicking images.