Discontinuity adaptive MRF model for remote sensing image analysis
P.C. Smits, Silvana Dellepiane · 2002
An approach is presented for the analysis of synthetic aperture radar (SAR) images that preserves better fine structures and borders in the image than classical methods. The method uses the discontinuity adaptive MRF label model combined with an observation model-based on a gamma distribution. This resulted in a new algorithm that is more suited to the segmentation of SAR images if one is interested in preserving details. Examples are given using real SAR data.