Markov Random Fields for SAR Remote Sensing Applications

This article aims at illustrating the powerfulness of Bayesian and specially Markovian frameworks for different remote sensing applications and in particular for SAR (Synthetic Aperture Radar) image processing. Indeed, the Markovian model is a very convenient way to introduce prior knowledge on the problem to solve. It will first be evoked with examples on the pixel level like filtering, segmentation and classification. Then higher level applications, like object recognition, and global image interpretation will be developed.