An Example of Application: Processing of Coherent Polarimetric Images

Abstract

In this chapter, we will illustrate the statistical image processing techniques described throughout this book on a nonconventional imaging modality: active polarimetric imagery. This will provide us with the opportunity of introducing important new aspects of the design of algorithms based on decision and estimation theory. In particular, we will analyze how to design computationally efficient detection and segmentation algorithms when the noise pdf does not belong to the exponential family. We will also present a general method for defining the contrast between two image regions in scalar or vectorial images with arbitrary statistics.