人工智能培训

搜索

深度学习论文:MS / HS融合网多光谱和高光谱图像融合(Multispectral and Hyperspectral Image Fusion by MS/

[复制链接]
~Limi多多~ 发表于 2019-1-11 11:07:19 | 显示全部楼层 |阅读模式
~Limi多多~ 2019-1-11 11:07:19 363 0 显示全部楼层
深度学习论文:MS / HS融合网多光谱和高光谱图像融合(Multispectral and Hyperspectral Image Fusion by MS/HS Fusion Net)与传统图像系统相比,高光谱成像有助于更好地了解不同材料的特性。然而,实际上通常只能以视频速率捕获高分辨率多光谱(HrMS)和低分辨率高光谱(LrHS)图像。在本文中,我们提出了一种基于模型的深度学习方法,用于合并HrMS和LrHSimage以生成高分辨率高光谱(HrHS)图像。具体而言,我们构建了一种新的MS / HS融合模型,该模型考虑了低分辨率图像的观测模型和沿着HSHS图像的光谱模式的低秩度知识。然后我们设计了一个迭代算法,通过利用近端梯度法求解模型。然后,通过展开设计的算法,我们构建了一个称为MS / HS融合网络的深度网络,通过卷积神经网络学习近端算子和模型参数。模拟和真实数据的实验结果证实了我们的方法在视觉和数量上的优越性,与这一研究领域的现有技术方法相比。
Hyperspectral imaging can help better understand the characteristics ofdifferent materials, compared with traditional image systems.However, onlyhigh-resolution multispectral (HrMS) and low-resolution hyperspectral (LrHS)images can generally be captured at video rate in practice.In this paper, wepropose a model-based deep learning approach for merging an HrMS and LrHSimages to generate a high-resolution hyperspectral (HrHS) image.In specific,we construct a novel MS/HS fusion model which takes the observation models oflow-resolution images and the low-rankness knowledge along the spectral mode ofHrHS image into consideration.Then we design an iterative algorithm to solvethe model by exploiting the proximal gradient method.And then, by unfoldingthe designed algorithm, we construct a deep network, called MS/HS Fusion Net,with learning the proximal operators and model parameters by convolutionalneural networks.Experimental results on simulated and real data substantiatethe superiority of our method both visually and quantitatively as compared withstate-of-the-art methods along this line of research.深度学习论文:MS / HS融合网多光谱和高光谱图像融合(Multispectral and Hyperspectral Image Fusion by MS/HS Fusion Net) WSLuowSWrbPoWwPc.jpg
URL地址:https://arxiv.org/abs/1901.03281     ----pdf下载地址:https://arxiv.org/pdf/1901.03281    ----深度学习论文:MS / HS融合网多光谱和高光谱图像融合(Multispectral and Hyperspectral Image Fusion by MS/HS Fusion Net)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

~Limi多多~当前离线
新手上路

查看:363 | 回复:0

快速回复 返回顶部 返回列表