SAR-to-Optical Image Translation via Thermodynamics-inspired Network

TL;DR

This paper introduces a thermodynamics-inspired neural network for SAR-to-optical image translation, improving image quality and structural preservation by modeling the translation process based on thermodynamic principles.

Contribution

The proposed S2O-TDN leverages thermodynamic theory to enhance SAR-to-optical translation, introducing a novel residual structure and a regularization branch for better feature learning and explainability.

Findings

Outperforms existing methods with more detailed textures

Achieves higher quantitative accuracy on the SEN1-2 dataset

Provides more stable and structure-preserving translations

Abstract

Synthetic aperture radar (SAR) is prevalent in the remote sensing field but is difficult to interpret in human visual perception. Recently, SAR-to-optical (S2O) image conversion methods have provided a prospective solution for interpretation. However, since there is a huge domain difference between optical and SAR images, they suffer from low image quality and geometric distortion in the produced optical images. Motivated by the analogy between pixels during the S2O image translation and molecules in a heat field, Thermodynamics-inspired Network for SAR-to-Optical Image Translation (S2O-TDN) is proposed in this paper. Specifically, we design a Third-order Finite Difference (TFD) residual structure in light of the TFD equation of thermodynamics, which allows us to efficiently extract inter-domain invariant features and facilitate the learning of the nonlinear translation mapping. In addition, we exploit the first law of thermodynamics (FLT) to devise an FLT-guided branch that promotes the state transition of the feature values from the unstable diffusion state to the stable one, aiming to regularize the feature diffusion and preserve image structures during S2O image translation. S2O-TDN follows an explicit design principle derived from thermodynamic theory and enjoys the advantage of explainability. Experiments on the public SEN1-2 dataset show the advantages of the proposed S2O-TDN over the current methods with more delicate textures and higher quantitative results.