SSA-GAN: Singular Spectrum Analysis-Enhanced Generative Adversarial Network for Multispectral Pansharpening
Lanfa Liu, Jinian Zhang, Baitao Zhou, Peilun Lyu, Zhanchuan Cai
Source abstract
Pansharpening is essential for remote sensing applications requiring high spatial and spectral resolution. In this paper, we propose a novel Singular Spectrum Analysis-Enhanced Generative Adversarial Network (SSA-GAN) for multispectral pansharpening. We designed SSA modules within the generator, enabling more effective extraction and utilization of spectral features. Additionally, we introduce Pareto optimization to the nonreference loss function to improve the overall performance. We conducted comparative experiments on two representative datasets, QuickBird and Gaofen-2 (GF-2). On the GF-2 dataset, the Peak Signal-to-Noise Ratio (PSNR) reached 30.045 and Quality with No Reference (QNR) achieved 0.920, while on the QuickBird dataset, PSNR and QNR were 24.262 and 0.817, respectively. These results indicate that the proposed method can generate high-quality pansharpened images with enhanced spatial and spectral resolution.
Evidence graph
No public relationships recorded yet.
Integrity note: This page is a factual metadata record created by deterministic ingestion. It is not a claim that the work moves a mathematical frontier or has been independently verified.