Feature Based Stitching of a Clear/Blurred Image Pair

Clear image stitching becomes mature now, but how to stitch blurred images with clear ones is still tough, mainly because of the ill-posed image deblurring. Existing deblurring methods assume spatial-invarint blur kernel, pixelwise constant blur kernel or planar camera motion. But these assumptions are difficult to hold in the real blurred image which is a spatial-variant blurred image by 3D camera motion. To overcome such limits for stitching blurred image, in this paper, we present a new feature based framework of stitching a clear/blurred image pair. It is inspired by recently proposed projetive warping models depicting the camera movement embedded in the capturing process of the blurred image. Our main contribution is a feature based algorithm for kernel estimation using the overlapped clear/blurred image pair. Experimental results demonstrate the effectiveness of the proposed stitching method.