Positioning and Mitigating Mining-Induced Phase Gradients for InSAR Phase Unwrapping

Interferometric Synthetic Aperture Radar (InSAR) has been widely used for deformation monitoring in mining areas. However, the nonlinear high-phase gradients caused by subsidence funnels are a major source of error in accurate measurements, as they present challenges for phase unwrapping. In this paper, we present a methodology to position and mitigate the mining-induced phase gradients. First, we utilize YOLOv10 to detect subsidence funnels that appear as small targets in SAR interferograms. Second, we model the nonlinear phase gradients by means of generalized Gaussian distribution, followed by minimizing the angular deviation between observed and modelled phase patterns in each detected interferometric patch. After defringing mining-induced phase gradients, we evaluate the impact of nonlinear high phase gradients on phase unwrapping, using synthetic data and Sentinel-1 dataset over Datong mining area, Shanxi Province. Compared to phase unwrapping without defringing, the proposed approach reduced the RMSE by 35.5% in the simulation. For the real data, the average number of unclosed pixels was reduced by 30.7% before and after defringing.