This work deals with the interferometric phase (IP) estimation from data collected by nonuniform multibaseline InSAR system. The approach proposed aims to recover the uniform linear array (ULA)data by means of a linear transformation of the data collected by the actual sensing geometry. After array interpolation, the IP estimation is carried out by means of the parametric spectral estimator root-MUSIC. Different InSAR scenarios are analized and the estimation accuracy is obtained performing a Monte Carlo analysis. Furthermore, the estimation performance is analyzed also in presence of array miscalibration and robust interpolation techniques are then proposed. As a benchmark, the hybrid Cramér-Rao bound is calculated in the case of interest.