Supervised Evolutionary Optimization of Bayesian Nonlocal Means Filter with Σ Preselection for Despeckling SAR Images

Conference: EUSAR 2012 - 9th European Conference on Synthetic Aperture Radar
04/23/2012 - 04/26/2012 at Nuremberg, Germany

Proceedings: EUSAR 2012

Pages: 4Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Gomez, Luis; Munteanu, Cristian (Electronic Engineering and Automatic Department, University of Las Palmas G.C., 35017, EITE, Campus Tafira, Spain)
Buemi, Maria; Berlles, Julio Jacobo; Mejail, Marta (Computation and Image Processing Group, University of Buenos Aires, Argentina)

Speckle reduction is an important problem in SAR image analysis. We present an interactive easy-to-use software package based on an evolutionary algorithm, to optimize a despeckling Bayesian nonlocal means filter with σ preselection for reducing the image variance while preserving the image mean. As a difference from other methodologies, there is an implication of the user, which provides a subjective validation to complete the Bayesian filter design. We apply the methodology on both synthetic and real intensity SAR images and the results show the effectiveness of the proposal.