OWA filters for SAR imagery

We investigate a subclass of ordered statistical filters for image processing, here called OWA filters. They are based on the OWA family of mean operators from the fuzzy sets literature. In particular, we are interested in the use of these filters to reduce speckle in SAR imagery. Moreover, we study how to use genetic algorithms in order to learn the weight vectors associated with the OWA operators. We also present an application on SAR imagery, simulated using Wishart distribution parameters estimated from a real-world application. We compare our results to those issued by two model-dependent SAR filters, as well as to some basic order statistical filters.