Accurate interpretation of interferometric image requires an extremely challenging task based on actual phase reconstruction for incomplete noise observation. In spite of the establishment of comprehensive solutions, until now, a guaranteed means of solution method is yet to exist. The initially observed interferometric image is formed by 2π-periodic phase image that wrapped within (-π, π]. Such inverse problem is further corrupted by noise distortion and leads to the degradation of interferometric image. In order to overcome this, an effective algorithm that enables noise suppression and absolute phase reconstruction of interferometric phase image is proposed. The proposed method incorporates an improved order statistical filter that is able to adjust or vary on its filtering rate by adapting to phase noise level of relevant interferometric image. Performance of proposed method is evaluated and compared with other existing phase estimation algorithms. The comparison is based on a series of computer simulated and real interferometric data images. The experiment results illustrate the effectiveness and competency of the proposed method.
UAVSAR observations of InSAR polarimetric phase diversity: implications for NISAR ionospheric phase estimation
