超光谱遥感图像特征分析  

Feature analysis of hyperspectral images

作　　者：刘恒殊[1] 彭风华[2] 黄廉卿[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130022 [2]广州市公安局,广东广州510435

出　　处：《光学精密工程》2001年第4期392-395,共4页Optics and Precision Engineering

摘　　要：超光谱遥感图像是一种三维立体图像，可反映地物的空间信息和光谱信息，其数据量庞大。本文通过对实际多光谱遥感图像的测试，分析了空间相关性和谱间相关性；计算出自相关系数和互相关系数；并与普通图像进行比较，揭示了超光谱图像的特征，为超光谱图像处理和压缩编码奠定了理论基础。The hyperspectral image is a three-dimensional image, which can provide both spatial information and spectral information. Because of the large size, it is difficult to be transmitted and restored. The paper describes the tests on hyperspectral images and analyzes the spatial correlation and spectral correlation. The self - correlation coefficient and cross-correlation coefficient are also calculated. In comparision with ordinary images, the feature of the hyperspectral images are revealed. The results obtained can be used in the research of remote sensing hyperspectral image compression.

关 键 词：超光谱图像 遥感图像 特征提取 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]