返回
Reverse degradation for remote sensing pan-sharpening
DOI:10.1016/j.jag.2026.105085.png)
摘要
En 中文
• Learning and reversing degradation for fusion via invertible neural networks. • Structural detail compensation enhances spatial feature representation. • Spatial–spectral contrastive learning improves alignment and fusion quality. • Interpretability efficacy coefficient quantifies model interpretability.
Keyword:
Pan-sharpening
Degradation model
Unsupervised fusion
Model interpretability
Self-learning
Contrastive learning
AI总结
对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。
期刊
IF:
8.6
论文数:
5.1K
被引数:
2.4W
机构
暂无机构信息

