Interferometric Phase Optimization Based on Total Power Polarization Optimization and Nonlocal Phase Linking

Interferometric phase optimization is a critical step in multitemporal interferometric synthetic aperture radar (MT-InSAR). The advent of multipolarimetric SAR satellites has boosted the polarimetric interferometric phase optimization techniques, where the polarimetric information can be used to enhance the interferometric phase quality. Polarimetric interferometric phase optimization approaches mainly include polarimetric coherence optimization (PCO) and polarimetric phase linking (PPL) algorithms. PCO algorithms that with good performance like exhaustive search polarimetric optimization (ESPO) are with high computational cost, especially for quad-polarimetric data cases. PPL algorithms mainly focus on distributed scatterers (DSs) and do not adopt the adaptive optimization for persistent scatterer (PS) pixels. In addition, their DS selection merely depends on the number of statistically homogeneous pixels (SHPs) through setting a threshold whose determination approach is usually not clearly defined. Moreover, the interferometric phase estimation results of PPL are affected by heterogeneous pixels. To overcome these limitations, by employing the total power (TP) polarization optimization and nonlocal phase linking (NL-PL), an adaptive polarimetric interferometric phase optimization algorithm for both PS and DS pixels is developed. The proposed algorithm has been evaluated using the simulated data, quad-polarimetric UAVSAR data, and dual-polarimetric Sentinel-1 data. The results demonstrate that in comparison with the previous methods, the proposed algorithm achieves superior interferometric phase quality, along with improved effectiveness and efficiency.