哈密土墩矿区高光谱影像蚀变矿物识别初步研究  

Application of hyperspectral imaging data to mineral identification in Tudun Cu-Ni deposit, Hami, Xinjiang

作　　者：周强[1] 甘甫平[2] 陈建平[1] 闫柏琨[1] 王润生[2] 

机构地区：[1]中国地质大学,北京100083 [2]中国国土资源航空物探遥感中心,北京100083

出　　处：《矿床地质》2006年第3期329-336,共8页Mineral Deposits

基　　金：本文得到国土资源部"百名优秀青年科技人才计划";国家自然科学基金(40201034);国土资源部科研项目(2002206)的联合资助

摘　　要：文章以哈密土墩矿区为例,利用航空高光谱HyMAP数据,从单矿物光谱与混合像元光谱之间的相似概率和模型化参数出发,发展了基于光谱混合组成的极大相关高光谱矿物识别方法,初步识别出该区与成矿关系密切的6种蚀变矿物及其空间分布趋势,并为野外地质工作与薄片鉴定所证实。该研究初步表明,高光谱遥感可为地质学中矿物的空间分布研究提供有用的新技术。Mineral is an important end-member in the processing of geological data. In this paper, a mineral identification method is put forward in the light of the similitude probability between pure mineral end-members and mixture pixel, including some model parameters obtained from mathematical and statistical means. Maximum coefficients based on pure mineral end-members and mixture pixel by the Spearman R and the Kendall tau statistical methods were designed and developed to extract the information of mineral with HyMAP data. Six minerals, namely, calcite, chlorite, epidote, serpentine, kaolinite and sericite, were identified by this method, with their spatial abundances given. The results were varified in the field and by the thin section identification. The hyperspectral imaging technique supplies a new and useful tool for mineral spatial analysis in geology.

关 键 词：地质学 HYMAP 矿物识别 数理统计 高光谱 土墩矿区 

分 类 号：P627[天文地球—地质矿产勘探]