Super-Resolution Algorithm Based on SNR and bination of Error-Parameter Analysis and Searching Method

, super-resolution image restoration approach is an ill-posed problem and its inversion is by regularization. It is important to set proper regularization parameter. Furthermore in ctical situations the blurring process is generally unknown or is known only to within a set ters, it is necessary to incorporate the blur identification into the restoration procedure. geometric warp matrixes are known, this paper presents a scheme in which regularization is estimated by Signal-Noise Ratio and error-parameter analysis is combined with method to identify blur function during super-resolution restoration. Simulation shows that heme the method of estimating regularization parameter can improve adaptive control ake convergence of restoration algorithm better during super-resolution image restoration, ively adjust this parameter when low-resolution images have different Signal-Noise Ratio; ination of error-parameter analysis and searching method can effectively identify blur which reduce the cost of computation and accelerate searching speed. : Super-resolution, Regularization Parameter, SNR, Blur Identification, Errorparameter Searching Method oduction olution (SR) image restoration is one of the most spot-lighted research areas, because it can the inherent resolution limitation of low-resolution (LR) imaging systems which can be ed and improve the performance of most digital image processing applications including aging, satellite imaging and video applications etc.. lly, SR image restoration approach is an ill-posed problem and its inversion is stabilized by tion. Bose et al. [1] pointed to the important role of regularization parameter and a CLS (constrained least square) SR reconstruction which generates the optimum value of rization parameter, using the L-curve method. super-resolution restoration algorithms, blurring process is assumed to be known. In many ituations, however, the blurring process is generally unknown or is known only to within a ameters, it is necessary to incorporate the blur identification into the restoration procedure. et al. [2] proposed a blind multi-channel SR algorithm by using multiple finite impulse