Improving the Spatial Resolution of GONG Solar Images to GST Quality Using Deep Learning

2025-10-06

AI Summary

To address the insufficient spatial resolution of GONG full-disk Hα images—which impedes clear delineation of fine solar structures such as filaments, prominences, and penumbrae—this paper proposes an enhanced super-resolution method based on Real-ESRGAN. The approach integrates Residual-in-Residual Dense Blocks (RRDB) with a relativistic discriminator and incorporates cross-modal image registration to reconstruct GONG low-resolution images into GST-level high-resolution outputs. Crucially, it preserves physical consistency while significantly improving structural fidelity. Quantitative evaluation yields MSE = 467.15, RMSE = 21.59, and correlation coefficient CC = 0.7794; qualitative assessment confirms visual quality closely approximating that of actual GST observations. This work represents the first systematic application of generative adversarial network-based super-resolution to solar Hα full-disk imagery, establishing a generalizable technical paradigm for high-fidelity restoration of low-resolution solar observational data.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Search and Optimization: Mixed Discrete/Continuous Search
• Machine Learning: Semi-Supervised Learning

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

High-resolution (HR) solar imaging is crucial for capturing fine-scale dynamic features such as filaments and fibrils. However, the spatial resolution of the full-disk H$α$ images is limited and insufficient to resolve these small-scale structures. To address this, we propose a GAN-based superresolution approach to enhance low-resolution (LR) full-disk H$α$ images from the Global Oscillation Network Group (GONG) to a quality comparable with HR observations from the Big Bear Solar Observatory/Goode Solar Telescope (BBSO/GST). We employ Real-ESRGAN with Residual-in-Residual Dense Blocks and a relativistic discriminator. We carefully aligned GONG-GST pairs. The model effectively recovers fine details within sunspot penumbrae and resolves fine details in filaments and fibrils, achieving an average mean squared error (MSE) of 467.15, root mean squared error (RMSE) of 21.59, and cross-correlation (CC) of 0.7794. Slight misalignments between image pairs limit quantitative performance, which we plan to address in future work alongside dataset expansion to further improve reconstruction quality.

Problem

• Enhancing GONG solar image resolution to GST quality
• Resolving fine-scale solar features like filaments and fibrils
• Addressing spatial resolution limitations in H-alpha solar imaging

Innovation

• GAN-based superresolution for solar image enhancement
• Real-ESRGAN with residual dense blocks
• Recovers fine details in sunspots and filaments