基于全波段反射光谱的花岗岩及其主要矿物自动识别研究--以康定某隧道为例  

Automatic identification of granite and its main minerals based on all-band reflection spectrum:A case study of a tunnel from Kangding area

作　　者：吕毓东 王世明 王代强 李磊[1] 裴秋明 LYU Yudong;WANG Shiming;WANG Daiqiang;LI Lei;PEI Qiuming(Faculty of Geosciences and Engineering,Southwest Jiaotong University,Chengdu 610097,Sichuan,China)

机构地区：[1]西南交通大学地球科学与工程学院,四川成都610097

出　　处：《矿产勘查》2024年第4期634-643,共10页Mineral Exploration

基　　金：四川省自然科学基金面上项目(2022NSFSC0410)资助。

摘　　要：岩矿的全波段光谱是高光谱遥感岩性解译的物理基础,也对岩矿的自动识别及分类具有重要意义。本文应用高性能便携地物光谱仪,在卤素灯全波段光源照射下,测量实验室、野外露头及隧道内岩矿样本的反射光谱,构建了94种常见岩石和矿物的用户地物波谱库,重点对典型花岗岩及构成花岗岩的6种主要矿物光谱特征进行研究。系统总结了花岗岩及其主要矿物波谱特征、主要吸收峰谷及产生的机理。运用光谱角匹配法(SAM)、光谱特征拟合法(SFF)和二进制编码分类法(BE)3种方法相结合并进行1∶1∶0.7权重分配的评分方式,将野外测得的8件中—酸性岩浆岩及矿物比对ENVI软件自带的JPL岩矿波谱库与用户地物波谱库,进行自动识别,结果显示,运用波谱匹配方法可实现岩矿自动识别和分类,用户地物波谱库在岩矿自动识别中具有明显优势。本研究以康定某隧道现场花岗岩及其主要矿物的全波段光谱研究为例,探讨了岩矿地物波谱在岩矿识别及自动分类中的价值,为岩矿快速、无损、便捷识别提供了新思路,建立的用户波谱库可应用于区域内岩矿的分类,对高光谱遥感岩性填图及矿产勘查具有较重要价值。The full-band spectrum of rocks and minerals is the physical basis for hyperspectral remote sensing lithology interpretation,and it is also of great significance for automatic identification and classification of rocks and minerals.In this contribution,a high-performance portable field spectrometer was used to measure the reflectance spectra of rock and mineral samples in laboratory,field outcrop,and tunnel under halogen lamp full-band light source irradiation.A user spectral database of 94 common rocks and minerals was constructed,focusing on the study of the spectral characteristics of typical granite and the six major minerals that make up granite.The spectral characteristics,major absorption peaks and valleys,and mechanisms produced for granite and its major minerals were systematically summarized.The spectral angle matching method(SAM),spectral feature fitting method(SFF),and binary encoding classification method(BE)were combined and scored using a 1∶1∶0.7 weighted allocation method.Eight pieces of acidic magmatic rocks and minerals measured in the field were compared with the JPL rock/mineral spectral database and user spectral database provided by ENVI software for automatic identification.The results showed that spectral matching methods can achieve automatic identification and classification of rocks and minerals,and the user spectral database shows obvious advantages in automatic identification of rocks and minerals.A case study of the full-band spectrum of granite and its main minerals in a tunnel site within the Kangding area discussed the value of rock and mineral spectral information in rock identifica⁃tion and automatic classification.It provides a new way to identify rocks and minerals quickly,non-destructively and conveniently.The established user spectral database can be applied to the classification of rocks and minerals in this area,which is of great value for hyperspectral remote sensing lithologic mapping and mineral exploration.

关 键 词：地物波谱仪 岩性识别 花岗岩 波谱匹配 遥感地质 

分 类 号：TP392[自动化与计算机技术—计算机应用技术]