| 光谱保真归一化卷积高光谱超分辨率重建方法 |
| |
| 引用本文： | 徐宏根,李春来,杨淼,董小环.光谱保真归一化卷积高光谱超分辨率重建方法[J].华南地质与矿产,2019,35(4):504-513. |
| |
| 作者姓名： | 徐宏根 李春来 杨淼 董小环 |
| |
| 作者单位： | 中国地质调查局武汉地质调查中心(中南地质科技创新中心),武汉,430205;中国科学院上海技术物理研究所,上海,200240;中国地质大学(武汉)信息工程学院,武汉,430074;中国人民解放军61618部队,北京,100088 |
| |
| 基金项目： | 民用航天预研项目;湖北省自然基金项目 |
| |
| 摘 要： | 高光谱遥感图像受到成像系统硬件限制,无法同时具有高空间分辨率和高光谱分辨率,较低的空间分辨率制约高光谱图像的应用,超分辨率重建技术可提高图像的空间分辨率。针对高光谱图像的超分辨率重建中光谱保真度的问题,在重建方法中耦合光谱保真度函数,结合结构自适应归一化卷积方法,提出基于光谱保真约束的归一化卷积方法。具体地,将图像局部邻域内像素间的光谱相关性作为约束条件,将与中心像素光谱类似的像素赋予较大权值,从而提高重建后图像光谱特性的保持程度。实验中分别从空间结构与光谱信息保真两个方面来评价重建后结果,结果表明该方法具有较好的光谱信息保持度。
|
| 关 键 词： | 超分辨率重建 归一化卷积 光谱保真度 结构自适应 |
A Super-resolution Reconstruction Method of Hyperspectral Image Considering Spectral Preservation Based on Normalized Convolution |
| |
| XU Hong-Gen,LI Chun-Lai,YANG Miao,DONG Xiao-Huan.A Super-resolution Reconstruction Method of Hyperspectral Image Considering Spectral Preservation Based on Normalized Convolution[J].Geology and Mineral Resources of South China,2019,35(4):504-513. |
| |
| Authors: | XU Hong-Gen LI Chun-Lai YANG Miao DONG Xiao-Huan |
| |
| Institution: | (Wuhan Center of China Geological Survey(Central South China Innovation Center for Geosciences),Wuhan 430205,Hubei,China;Shanghai Institute of Technical Physics,Chinese Academy of Sciences,Shanghai 200240,China;Faculty of Information Engineering,China University of Geoscience(Wuhan),Wuhan 430079,Hubei,China;The Chinese People's Liberation Army 61618,Beijing 100088,China) |
| |
| Abstract: | Hyperspectral remote sensing images are characterized by having a low spatial resolution and a high spectral resolution, which restricts the application of hyperspectral images. Super-resolution Reconstruction is a technique to increase images spatial resolution. A super-resolution reconstruction method of hyperspectral image considering spectral preservation is proposed by normalized convolution, SPNC for short, which is centered on spectral fidelity. Specifically, the spectral correlation is applied in the neighborhood as constraint conditions, which designs spectral preserving function to improvethe spectrum characteristics of reconstructed image. In the experiment, the reconstructed results were evaluated from two aspects: spatial structure and spectral information fidelity. The results indicate that the method has better spectral information retention. |
| |
| Keywords: | super-resolution reconstruction normalized convolution spectral preservation structure-adaptive |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
