主成分分析在遥感影像数据中的实例应用

主成分分析(Principal Component Analysis)是根据变量之间的相互关系,尽可能不丢失信息地用几个综合性指标表示多个变量的方法。在多(高)光谱图像中,由于各波段的数据间具有相关性,因此包含许多冗余信息。通过主成分分析法可以把遥感图像中所含的大部分信息用少数波段表示出来,这样就可以几乎不丢失数据但可以减少数据量,消除冗余信息。在遥感数据处理时用主成分分析法作数据分析前的预处理,以达到数据压缩和图像增强的效果,更加有利于影像信息提取。文章对主成分分析在遥感图像处理中的实际应用进行了实例示范应用研究。

Practical Application of Principal Component Analysis in Remote Sensing Image Processing

Principal component analysis is a method to express several multivariable factors without losing information as far as possible based on relations of variables. In multispectral images and hyperspectrum images, there is much redundant information because of relaion of the data in different bands, including many redundant information. Through principal component analysis, most information of remote sensing images can be expressed with fewer bands. Thus, it can not only reduce data size, but also eliminate redundant information. Principal component analysis method is always used to conduct data preprocessing in order to depress data and enhance images. In this paper, practical application of principal component analysis in remote sensing image processing has been studied.