Topics In Super-resolution Image Reconstruction

Confined by the available physical conditions, imaging systems produce images always with a limited resolution. However, images with higher resolution are more desired for application. Super-resolution techniques offer a possibility to produce an image with a higher resolution from a set of images with lower resolution. The underlying mechanism is implied in the fact that different sub-pixel displacement of each low resolution images contains different information of the high resolution image.The pioneer work of super-resolution reconstruction may go back to 1984 by Tsai and Huang. Since then, many researchers have devoted themselves to the work in this area. Currently, the research is focused to a few points, such as high precision sub-resolution registration algorithm; blind super-resolution methods; robust and efficient reconstruction; real-time processing techniques.The main work of this paper is as follows:(1) A registration algorithm method based on the interpolation and hierarchical block-match on the low resolution image dimension is drawn, which utilizes the relation between the sub-pixel image value with integral pixel image value, and the hierarchical block-match and FFT algorithm are adopted to optimize the registration parameters. The registration algorithm is characteristic of high precision and has lower computation complexity than traditional Taylor series expansion method.(2) Super-resolution reconstruction is an ill-posed problem, an efficient computing algorithm is drawn on the Tikhonov regularization. The algorithm is a non-iterative, which utilizes the FFT algorithm and the inverse of little-size matrices. The algorithm achieves de-convolution and fusion from the low-resolution images simultaneity, and almost performs in real-time.(3) For the reconstruction with Tikhonov regularization, the precision estimation for the regularization parameter could improve the reconstruction image quality. In this paper, the...