UnwrapDiff: A Conditional Diffusion Model for InSAR Phase Unwrapping

TL;DR

UnwrapDiff introduces a diffusion model-based framework for InSAR phase unwrapping that incorporates traditional algorithms as guidance, improving robustness and accuracy in noisy, real-world scenarios.

Contribution

The paper presents a novel DDPM-based method that integrates SNAPHU outputs as conditional guidance, enhancing phase unwrapping performance over existing approaches.

Findings

Achieves 10.11% reduction in NRMSE compared to SNAPHU.

Improves reconstruction quality in challenging cases like dyke intrusions.

Demonstrates robustness against atmospheric effects and diverse noise patterns.

Abstract

Phase unwrapping is a fundamental problem in InSAR data processing, supporting geophysical applications such as deformation monitoring and hazard assessment. Its reliability is limited by noise and decorrelation in radar acquisitions, which makes accurate reconstruction of the deformation signal challenging. We propose a denoising diffusion probabilistic model (DDPM)-based framework for InSAR phase unwrapping, UnwrapDiff, in which the output of the traditional minimum cost flow algorithm (SNAPHU) is incorporated as conditional guidance. To evaluate robustness, we construct a synthetic dataset that incorporates atmospheric effects and diverse noise patterns, representative of realistic InSAR observations. Experiments show that the proposed model leverages the conditional prior while reducing the effect of diverse noise patterns, achieving on average a 10.11\% reduction in NRMSE compared to SNAPHU. It also achieves better reconstruction quality in difficult cases such as dyke intrusions.