Empirical Bayesian Estimation of the Interferometric SAR Coherence Magnitude

SAR interferometry has developed rapidly in recent years and now allows measurements of subtle deformation the Earth's surface with millimeter accuracy. All state-of-the-art processing methods require a precise coherence estimate. However, this estimate is random variable biased toward higher values. Up to now, little published on Bayesian estimation degree coherence. The objective paper develop empirical estimators assess their characteristics by simulations. understood as regularization maximum likelihood estimation. more information used stricter general prior, accurate will be. Three levels prior are developed: (1) an uninformative (2) informative which can be implemented (2a) less strict (2b) prior. priors described single parameter only i.e. underlaying reports bias, standard deviation root mean square error (RMSE) estimators. It was found that all Bayes improve from small samples for coherences compared conventional sample estimator. E.g. zero estimated expected posteriori estimator without additional 33.3% reduced bias using three only. Assuming 0.6, 51.3% 36.6% In addition, it work very well even extremely size 2