Research on a maximum likelihood method for space-variant coded image deblurring

Coded imaging is an important technique in high energy physics, astronomy and many other areas. However, popular image reconstruction methods are not ideal for space-invariant systems, which badly restricts their application. In this paper, we propose a method based on the maximum-likelihood method that performs image reconstruction on coded images with general space-variant degradation. A universal space-variant model is adopted, and the maximum-likelihood method is applied to a Gaussian stochastic noise model to derive an iterative deblurring formula. In practice, this method shows fine smoothness, convergence and good performance for strongly space-variant reconstructions.