Temporal Integration Method for Improving Spatial Resolution of Moving Images

Abstract

This paper presents a method to acquire high resolution images by integrating corresponded pixels over a series of images. To estimate accurate correspondence between the pixels in the two temporally serial image, we propose a new block matching method applicable to the motion with affin transformation. The method employs an iterative technique which analysis blocks in the image is transformed toward decreasing mean square error(MSE) in the block, and thus enable analysis blocks to track over more frames compared to usual method. In this paper, we apply this method to a series of 50 frames and improved the resolution in a image.