Multichannel SAR Interferometry based on Statistical Signal Processing

Interferometric SAR systems allow to reconstruct height profile of earth surfaces. The height reconstruction is based on phase unwrapping operation, which is an ill-posed problem since it admits infinite number of solutions. The phase unwrapping problem can be solved using the multichannel InSAR approach, based on the combination of different images of the scene obtained by slightly different positions. An effective way to combine the different interferograms is using statistical estimation theory. In particular the Maximum Likelihood and Maximum a Posteriori multichannel approach have proved to be effective and to be able to restore the uniqueness of the solution. In this thesis the statistical multichannel phase unwrapping is deeply analyzed. In particular, instruments and methods to use InSAR multichannel configuration on real data and in urban areas are provided. Moreover, a new fast and efficient multichannel phase unwrapping algorithm is provided. The presented innovative techniques have been tested on simulated data and real data, showing the consistence and effectiveness of the proposed approaches.