Robust nonlinear adaptive image restoration in land mine detection problem

To enhance image quality, deconvolution is often used as an alternative to additional measurements. This relates to solution of an inverse ill-posed problem, and consists in mathematical compensation for the degradation using image restoration and noise suppression methods. The aim of this paper is to demonstrate the advantages of the proposed robust estimation strategy in the restoration of low-contrast radiometry images, and its possibilities regarding mine detection in an environment of objects with similar form and dimensions. The paper presents a robust approach to image restoration that combines the properties of classical regularized iterative algorithms and robust features based on the concept of M-estimators. The proposed technique could be efficiently used for the solution of the depth resolution enhancement problem in radar applications.