A method for optimal linear super-resolution image restoration

In this paper, we propose a super-resolution (pixel grid refinement) method for digital images. It is based on the linear filtering of zero-padded discrete signal. We introduce continuous-discrete observation model to create reconstruction system. The proposed typical real-world imaging systems - an initially continuous signal first undergoes (dynamic) distortions and then subjected sampling effect additive noise. optimal in sense mean square recovery error minimization. theoretical part article, general scheme presented expressions impulse frequency responses system are derived. An expression such restoration also For sake brevity, entire description one-dimensional signals, but obtained results can easily be generalized case two-dimensional experimental section paper devoted analysis depending parameters model. significant superiority terms demonstrated comparison with interpolation, which usually used refine image pixels.