Advancing InSAR Shift Measurement: Refining Precision and Phase Unwrapping Performance Analysis of SSENet

Interferometric Synthetic Aperture Radar (InSAR) shift measurement plays a key role in image coregistration and absolute phase measurement and has significant applications in the InSAR processing workflow. However, the current shift measurement algorithms are limited by the relative bandwidth of the SAR system, resulting in low resolution and accuracy. SSENet is a recent InSAR shift measurement approach that utilizes deep learning to address these issues to some extent. This paper proposes a calibration method for SSENet, which introduces a lightweight neural network designed to refine the marginally biased output shifts. Furthermore, we demonstrate the performance of the refined SSENet algorithm and its potential in assisting phase unwrapping using LSAR-01 bistatic Synthetic Aperture Radar (SAR) data.