arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2604.16841cs.CVcs.LG
收藏

当地球基础模型遇见扩散:一种用于地表温度超分辨率的应用

When Earth Foundation Models Meet Diffusion: An Application to Land Surface Temperature Super-Resolution

University of Alabama(阿拉巴马大学);Emory University(埃默里大学);University of Michigan(密歇根大学);University of South Florida(佛罗里达州立大学);New York University(纽约大学)

Yiheng Chen, Zihui Ma, Peishi Jiang, Yilong Dai, Qikai Hu, Xinyue Ye, Lingyao Li, Rita Sousa, Runlong Yu

AI总结:

本文提出EFDiff框架,利用地球基础模型指导扩散模型进行极端空间退化下的超分辨率重建,通过交叉注意力机制提升细尺度重建效果,优于基线方法。

摘要:

地表温度(LST)超分辨率对于环境监测至关重要。然而,粗粒度热观测严重限制了细尺度结构的确定性。本文提出地球基础模型引导扩散(EFDiff)框架,利用Prithvi-EO-2.0地球基础模型将高分辨率多光谱反射率编码为地理空间嵌入,通过交叉注意力机制注入去噪网络,指导从高度退化的观测中重建细尺度结构。我们研究了两种变体EFDiff-ε和EFDiff-x₀,提供了感知真实性和像素级保真度之间的互补权衡。在极端32倍尺度差距下,使用包含242,416个共注册的Landsat热反射率块的全球多样化基准评估EFDiff。结果表明,EFDiff在感知真实性和像素级保真度之间取得了平衡,优于基线方法。尽管本文在LST超分辨率背景下提出EFDiff,该框架广泛适用于可以利用预训练地理空间表示指导生成重建的遥感问题。

英文摘要:

Land surface temperature (LST) super-resolution is important for environmental monitoring. However, it remains challenging as coarse thermal observations severely underdetermine fine-scale structure. In this paper, we propose Earth Foundation Model-guided Diffusion (EFDiff), a novel framework for super-resolution under extreme spatial degradation. EFDiff uses the Prithvi-EO-2.0 Earth foundation model to encode high-resolution multispectral reflectance into geospatial embeddings, which are injected into the denoising network via cross-attention to guide fine-scale reconstruction from highly degraded observations. We study two variants, EFDiff-$ε$ and EFDiff-$x_0$, which offer complementary trade-offs between perceptual realism and pixel-level fidelity. We evaluate EFDiff under an extreme $32\times$ scale gap using a globally diverse benchmark comprising 242,416 co-registered Landsat thermal-reflectance patches. Results show that EFDiff consistently outperforms baseline methods and that cross-attention conditioning by EFM is more effective than HLS channel concatenation. Although we present EFDiff in the context of LST super-resolution, the framework is broadly applicable to remote sensing problems in which pretrained geospatial representations can guide generative reconstruction.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑