Favoritar este trabalho

We compare WM, OWA and WOWA filters in mean SAR images, respectively based on the parametrized families of operators WM (Weighted Means), OWA (Ordered Weighted Average) and WOWA (Weighted OWA). The WM and OWA operators use a single vector to weight an input data vector: one whose weights correspond to the same positions in the input, and another that considers the ordered positions of the input, respectively. WOWA operators, however, use both of these weight vectors. We present an application on SAR imagery, using simulated images derived from from a real-world scene and 5x5 windows, in which we make use of a Genetic Algorithm to tune the parameters of these families of operators.