Robust cutting plane pure integer programming phase unwrapping algorithm considering fringe frequency for dual-baseline InSAR

Multi-baseline (MB) phase unwrapping (PU) represents the core processing step of MB InSAR, which overcomes discontinuous terrain height estimation. However, MBPU faces the challenge of low noise robustness. A new robust cutting plane–pure integer programming (CP-PIP) PU algorithm that considers the fringe frequency for dual-baseline InSAR is proposed to address this problem. First, we establish a closed-region PIP model with one objective function and two constraints using prior information about the fringe frequency and the relationship between the interferometric phase difference and the same relative elevation. Then, after determining secant equations, we begin bidirectional traversal with the integer solution of one ambiguity number variable to verify whether another ambiguity number variable satisfies the integer condition using the CP-PIP method. Finally, two-neighborhood phase gradients in both the azimuth and range directions are introduced to extract abrupt topographic change points, and the ambiguity number of the mis-unwrapping point at the center of the window function is replaced by that with the highest frequency to complete the PU. Comprehensive comparisons using both simulated data and real data confirm the effectiveness, universality, and reliability of our approach. Compared with the most effective minimum-cost flow (MCF) algorithm in single-baseline (SB) PU and the two-stage programming approach (TSPA) in MBPU, the proposed algorithm not only achieves better and more stable unwrapping in the region of abrupt phase change and dense interferometric fringes but also improves the unwrapping accuracy. The root mean square error of the proposed algorithm is less than those of MCF and TSPA algorithms by more than 60% and 10%, respectively. The full implementation of CP-PIP is publicly available https://github.com/Shiji-Yang/MBPU.git.