Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Physics Encoded Spatial and Temporal Generative Adversarial Network for Tropical Cyclone Image Super-resolution

Ruoyi Zhang, Jiawei Yuan, Lujia Ye, Runling Yu, Liling Zhao

TL;DR

A disentangled generator architecture incorporating a PhyCell module, which approximates the vorticity equation via constrained convolutions and encodes the resulting approximate physical dynamics as implicit latent representations to separate physical dynamics from visual textures, and a dual-discriminator framework is introduced.

Abstract

High-resolution satellite imagery is indispensable for tracking the genesis, intensification, and trajectory of tropical cyclones (TCs). However, existing deep learning-based super-resolution (SR) methods often treat satellite image sequences as generic videos, neglecting the underlying atmospheric physical laws governing cloud motion. To address this, we propose a Physics Encoded Spatial and Temporal Generative Adversarial Network (PESTGAN) for TC image super-resolution. Specifically, we design a disentangled generator architecture incorporating a PhyCell module, which approximates the vorticity equation via constrained convolutions and encodes the resulting approximate physical dynamics as implicit latent representations to separate physical dynamics from visual textures. Furthermore, a dual-discriminator framework is introduced, employing a temporal discriminator to enforce motion consistency alongside spatial realism. Experiments on the Digital Typhoon dataset for 4$\times$ upscaling demonstrate that PESTGAN establishes a better performance in structural fidelity and perceptual quality. While maintaining competitive pixel-wise accuracy compared to existing approaches, our method significantly excels in reconstructing meteorologically plausible cloud structures with superior physical fidelity.

Physics Encoded Spatial and Temporal Generative Adversarial Network for Tropical Cyclone Image Super-resolution

TL;DR

A disentangled generator architecture incorporating a PhyCell module, which approximates the vorticity equation via constrained convolutions and encodes the resulting approximate physical dynamics as implicit latent representations to separate physical dynamics from visual textures, and a dual-discriminator framework is introduced.

Abstract

High-resolution satellite imagery is indispensable for tracking the genesis, intensification, and trajectory of tropical cyclones (TCs). However, existing deep learning-based super-resolution (SR) methods often treat satellite image sequences as generic videos, neglecting the underlying atmospheric physical laws governing cloud motion. To address this, we propose a Physics Encoded Spatial and Temporal Generative Adversarial Network (PESTGAN) for TC image super-resolution. Specifically, we design a disentangled generator architecture incorporating a PhyCell module, which approximates the vorticity equation via constrained convolutions and encodes the resulting approximate physical dynamics as implicit latent representations to separate physical dynamics from visual textures. Furthermore, a dual-discriminator framework is introduced, employing a temporal discriminator to enforce motion consistency alongside spatial realism. Experiments on the Digital Typhoon dataset for 4 upscaling demonstrate that PESTGAN establishes a better performance in structural fidelity and perceptual quality. While maintaining competitive pixel-wise accuracy compared to existing approaches, our method significantly excels in reconstructing meteorologically plausible cloud structures with superior physical fidelity.
Paper Structure (16 sections, 7 equations, 4 figures, 2 tables)

This paper contains 16 sections, 7 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Overall architecture
  4. Physics encoded generator
  5. Disentangled architecture
  6. Phycell injection
  7. Dual discriminators
  8. Spatial discriminator
  9. Temporal discriminator
  10. Loss Function
  11. Experiments
  12. Dataset and implementation details
  13. Comparison study
  14. Ablation Study
  15. Conclusion
  16. ...and 1 more sections

Figures (4)

  • Figure 1: Overall architecture of PESTGAN, the LR pictures are first upsampled and then sent into the physics-encoded generator. After generation, the SR pictures will be sent into a dual-discriminator framework, which is designed for spatial fidelity and temporal coherence.
  • Figure 2: The structure of PhyCell
  • Figure 3: Visual comparison of the super-resolution results of different models on the 4× super-resolution task across various test sequences.
  • Figure 4: Visual comparison of the super-resolution results of ablation study on the 4× super-resolution.