基于实测波谱标识的Hyperion数据岩性识别方法

A study of the lithologic identification based on Hyperion and field spectra

摘要: 以中甸普朗典型斑岩铜矿区和松诺典型矿点的2种主要含矿岩性为实验研究对象，将特征岩性实测波谱数据转换为Hyperion影像像元数据，并利用MNF变换对高光谱数据进行降噪和降维，以第一个MNF变换主成分作为分析数据，以所构建像元的亮度值为参考分割点，通过密度分割，进行岩性识别，并通过与已有的地质资料的对比分析和野外调查，验证结果的精度。实验结果证明，在影像处理过程中有机地融入地面实测波谱，可较有效地识别中甸地区与矿相关的石英二长斑岩和石英闪长玢岩。

Abstract: A method for lithologic identification based on Hyperion data was developed and tested at Pulang and Songnuo of Zhongdian, Yunnan Province. Field spectra were turned into pixel of Hyperion. In this study, MNF was applied to Hyperion bands covering SWIR band and MNF band1 was used to identify lithology by density slicing. The mapping results were evaluated by the comparison with the geological maps. The results show that the method of turning field spectra into Hyperion pixel can effectively detect quartz diorite porphyrite and beschtauite in Zhongdian.