Bayesian model based city reconstruction from high resolution ISAR data

We present Bayesian information extraction methods for feature extraction from metric resolution interferometric radar data. The extracted information is naturally related to a three-dimensional environment. We consider this information for the reconstruction of complex urban settlement areas. We target the higher complexity of the scene that is captured by the high resolution of the sensors by using hierarchical models of the data. Their stochastic nature allows us to deal with the uncertainties that are inherent in both the acquisition process and in the data themselves. We present results of the application of the described concept to the separation and characterization of typical man-made objects appearing in urban areas.