SNR and DFD Based Blind Image Super-Resolution

摘要

In this paper, a kind of blind image super-resolution algorithm is proposed, in which Error-Parameter-Analysis (EPA) is combined with the searching method to identify parameterized blur function and a regularized super-resolution (SR) image restoration algorithm simultaneously considering signal-to-noise ratio (SNR) and displaced frame differential (DFD) is proposed to restore high-resolution (HR) image. The adoption of the searching method shrinks identification period. Simulation results show that the proposed blind SR algorithm can obtain approximate blur function and restore well-visual HR image.