Beyond Backscatter: InSAR coherence from detected SAR images
Researchers have developed a deep learning framework capable of estimating InSAR coherence directly from detected SAR images, eliminating the need for precise coregistration. A Residual U-Net model was trained on Sentinel-1 data to learn the relationship between backscatter magnitudes and coherence. This approach demonstrates improved accuracy over existing intensity-based methods and shows strong generalization across various locations and temporal baselines, enabling large-scale applications. AI
IMPACT Enables large-scale application of InSAR coherence estimation for mission design, change monitoring, and mapping tasks.