arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.02552eess.IVcs.CV

利用合成丰度数据的无监督超分辨率处理遥感超光谱图像

Super-résolution non supervisée d'images hyperspectrales de télédétection utilisant un entraînement entièrement synthétique

LTCI, Télécom Paris, Institut Polytechnique de Paris(LTCI,Telecom Paris,巴黎高等理工学院)

Xinxin Xu, Yann Gousseau, Christophe Kervazo, Saïd Ladjal

AI总结:

本文提出一种基于合成数据的无监督方法,用于提升遥感超光谱图像的空间分辨率,通过解混和神经网络训练实现有效重建。

摘要:

超光谱单图像超分辨率(SISR)旨在在提升空间分辨率的同时保持超光谱图像丰富的光谱信息。现有大多数方法依赖于高分辨率地面真实数据的监督学习,但实践中这种数据往往不可用。为克服这一限制,我们提出了一种基于合成丰度数据的无监督学习方法。首先通过超光谱解混将超光谱图像分解为端成员和丰度图。然后利用死叶模型生成的数据训练神经网络,以超分辨率方式处理这些图。最终通过将超分辨率后的丰度图与端成员重新组合来重建超分辨率的超光谱图像。实验结果证明了我们方法的有效性以及合成数据对训练的相关性。

英文摘要:

Hyperspectral single image super-resolution (SISR) aims to enhance spatial resolution while preserving the rich spectral information of hyperspectral images. Most existing methods rely on supervised learning with high-resolution ground truth data, which is often unavailable in practice. To overcome this limitation, we propose an unsupervised learning approach based on synthetic abundance data. The hyperspectral image is first decomposed into endmembers and abundance maps through hyperspectral unmixing. A neural network is then trained to super-resolve these maps using data generated with the dead leaves model, which replicates the statistical properties of real abundances. The final super-resolution hyperspectral image is reconstructed by recombining the super-resolved abundance maps with the endmembers. Experimental results demonstrate the effectiveness of our method and the relevance of synthetic data for training.

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。

↑