基于高光谱对花岗岩中长石含量测定研究  

Study on Determination of Feldspar Content in Granite Based on Hyperspectral

作　　者：彭逸飞 周可法[1,3] 王金林 王珊珊[1,3] 周曙光 白泳[1,2,3] Peng Yifei;Zhou Kefa;Wang Jinlin;Wang Shanshan;Zhou Shuguang;Bai Yong(Xinjiang Research Center for Mineral Resources,Xinjiang Institute of Ecology and Geography,Chinese Academy of Sciences,Urumqi,Xinjiang,830011,China;University of Chinese Academy of Sciences,Beijing,100049,China;Xinjiang Key Laboratory of Mineral Resources and Digital Geology,Urumqi,Xinjiang,830011,China)

机构地区：[1]中国科学院新疆生态与地理研究所新疆矿产资源研究中心,新疆乌鲁木齐830011 [2]中国科学院大学,北京100049 [3]新疆矿产资源与数字地质重点实验室,新疆乌鲁木齐830011

出　　处：《新疆地质》2023年第2期291-295,共5页Xinjiang Geology

基　　金：自治区人才专项计划-天山优秀青年(2019Q033);新疆维吾尔自治区重大科技专项(2021A03001-3);国家自然科学基金-新疆联合基金重点项目(U1803241);中国科学院重点领域部署项目(ZDRW-ZS-2020-4-2)共同资助。

摘　　要：长石是地表岩石最主要的造岩矿物。为研究光谱分辨率及颗粒尺度对长石含量预测的影响,测定由石英、钾长石及斜长石粉末混合成的模拟花岗岩光谱信息。采用高斯响应函数重采样法,进行多光谱变换后,使用随机森林、支持向量机及极端梯度提升建模,得到钾长石及斜长石含量反演的最佳光谱宽度分别是10 nm和30 nm。依次去除矿物粒径较小样本,建模反演可得斜长石和钾长石的平均决定系数(M),分别增加0.25与0.26,表明岩石中矿物颗粒尺度越大,矿物反演精度越高。研究结论可得,基于可见光-近红外波段信息反演岩石中矿物含量。据获得的长石含量反演最佳波段宽度及颗粒尺度,为设计合适高光谱传感器进行大范围地表岩石组分检测提供参考。Feldspar is the main rock forming mineral of surface rocks.In order to study the influence of spectral resolu-tion and particle size on the prediction of feldspar content,the spectral information of simulated granite mixed with quartz,potassium feldspar and plagioclase powder was measured.Gaussian response function resampling method is used for multispectral transformation,and random forest,support vector machine and extreme gradient boosting are used for modeling.It is obtained that the optimal spectral widths of potassium feldspar and plagioclase content inversion are 10 nm and 30 nm respectively.In addition,in order to remove samples with smaller particle size,the average coefficient of determination(M)of plagioclase and K-feldspar increases by 0.25 and 0.26 respectively,indicating that the larger the mineral particle size is,the higher the accuracy of mineral inversion is.It can be concluded that the content of minerals in rocks can be retrieved based on the visible near-infrared band information,and the optimal band widths and particle size can be retrieved according to the obtained feldspar content,which provides a reference for designing a suitable hyper-spectral sensor for large-scale detection of surface rock components.

关 键 词：长石 波段宽度 颗粒尺度 含量反演 高光谱 

分 类 号：P619.235[天文地球—矿床学]