包络线消除法及其在野外光谱分类中的应用

Continuum Removal and Its Application to the Spectrum Classification of Field Object

根据云南省鹤庆县北衙金矿区野外地物(岩石和植被)的高光谱数据,利用VC抖实现包络线消除算法,运用 对应分析和聚类分析方法,处理包络线消除前后数据。对比认为包络线消除法归一化后的光谱曲线能突出其吸收 特征,减弱一些野外因素的影响,可更好地进行分类,并得到该地区用于野外识别岩石的波段范围。

In this paper,hyperspectral data of rock and vegetaticm are s urveyed in Beiya Gold mine, Heqing County, Yunnan Province,and are used to obtain information. With the aid of a VC＋＋ program,the continuum removal is implemented. At the same time, correspondence analysis and duster analysis are used to process original hyperspectral data of rock and vegetation, and their eorresponding normalized data, respectively. The difference in results of normalized data and the original data processed by two methods shows that normalized spectrum curve enhances spectral absorption features greatly, and the effect of factors in the field is decreased obviously,leading to a better classifica.tion of objects. On the other hand, the spectrum ranges are obtained to identify rock and vegetation in the field,and provide a due to hyperspectral image of objects.