Parameter Analysis of Conditional Morphological Operators for image restoration

Conditional morphology is a nonlinear method for simultaneous edge detection and image smoothing. It is derived by defining statistical models for the input-output relation as in statistical morphology, with the introduction of a line process taking into account the discontinuities between different regions of the image. It results to be a method presenting the advantages of both morphological (simplicity) and Bayesian (robustness) methods and it allows one to perform image restoration by applying a continuation method. This is performed applying a scheduling of the temperature parameters introduced with the definition of statistical models. In this paper, different kinds of temperature scheduling are analyzed. Experimental results are shown from processing of Synthetic Aperture Radar (SAR) images.