Unwrapped Phase Estimation via Normalized Probability Density Function for Multibaseline InSAR

Interferometric synthetic aperture radar (InSAR) is a powerful technique for obtaining terrain information based on the interferometric phase. Multibaseline (MB) InSAR is an extension of the conventional InSAR and is used to improve the estimation accuracy and reliability of the unwrapped phase. Based on a newly defined normalized phase probability density function (pdf), a novel wrapped-to-unwrapped phase (W2UP) estimation method is proposed for MB-InSAR. First, the concept of the normalized pdf is introduced to overcome the limitation of the fixed
$2\pi $
period for different baseline cases. Then, a new maximum likelihood estimation is established using the MB normalized pdfs, which has a much steeper peak around the true phase value than the single baseline case and leads to higher estimation accuracy. The proposed W2UP method estimates the unwrapped phase from multiple filtered interferograms, so it is less influenced by the phase noise. Both the theoretical analysis and results using the simulated and real MB data are provided to verify the effectiveness of the proposed method.