Robust edge-preserving image restoration in the presence of non-Gaussian noise

An approach to image restoration is presented which combines the properties of classical regularised iterative algorithms and M-estimation. The method is based on a penalised maximum likelihood estimation incorporating a generalised robust objective function which takes into account non-Gaussian noise and edge-preserving image priors. Results are presented which demonstrate the effectiveness of the method with low-resolution noisy images of simulated landmines.