摘要
在遥感图像检索中,光谱特征的应用最为广泛。本文研究了基于光谱特征进行遥感图像检索的方法。针对目前应用越来越广泛的多光谱、高光谱遥感图像波段多的特点,提出了基于K-L变换的检索方法,将多维图像降维处理,在此基础上提取遥感图像的光谱特征,通过检索图像与目标图像的光谱特征对比实现多光谱遥感图像的检索,并通过实验验证了本文方法的有效性。
The spectral character is the basic character in remote sensing images, and is naturally being used the most widely in remote sensing image retrieval. Considering the characteristic of multi-spectral or hyper-spectral remote sensing images, K-L transformation based image retrieval method is proposed in this paper. Instead of original multi-spectral image, image retrieval can be carried out using the histogram of the first three principal component images; the distance of the remote sensing images can be calculated after extracting the spectral character. This is also proved to be effective and time saving by the experiments in this paper.
出处
《北京测绘》
2012年第5期56-59,共4页
Beijing Surveying and Mapping
关键词
摄影测量与遥感
遥感图像检索
光谱特征
多光谱图像
K-L变换
photogrammetry and remote sensing
remote sensing image retrieval
spectral character
multi-spectral images
K-L transformation
