基于光谱匹配的热红外高光谱数据岩性分类研究

发布时间：2018-01-02 21:14

  本文关键词：基于光谱匹配的热红外高光谱数据岩性分类研究　出处：《现代地质》2016年01期 　论文类型：期刊论文

  更多相关文章： 热红外高光谱 光谱特征 光谱离散能级波形匹配法 甘肃柳园地区 岩性分类

【摘要】：从岩石光谱出发,结合光谱谱带强度特征和光谱波形特征,针对机载热红外高光谱数据(TASI),在以往算法基础上,提出一种改进的算法——光谱离散能级波形匹配法(SDEM),并将其运用到岩性分类研究中。SDEM算法能识别岩石光谱间的微小差异,并在充分考虑光谱谱带强度和波形特征的同时,有效减弱数据噪声。与传统的岩性分类方法——高光谱角度制图法(SAM)相比,改进的算法能更精确地区分岩石相似光谱,识别易混淆岩性,对出现"异物同谱"现象的岩石也具有更好的区分能力。将SDEM、SAM方法应用于甘肃柳园地区TASI数据岩性分类研究中,可看出SDEM方法能识别出SAM未识别或识别错误的岩性。通过研究区野外查证,可知SDEM方法所得岩性分类结果更符合岩石实际分布情况。可见光谱离散能级波形匹配法具有较好的岩性分类效果,能更好地区分地物。
[Abstract]:Based on the rock spectrum, combined with the spectral band intensity and spectral waveform characteristics, the airborne thermal infrared hyperspectral data are analyzed on the basis of previous algorithms. An improved algorithm, the spectral discrete energy level waveform matching method, is proposed and applied to the lithologic classification study. The SDEM algorithm can identify the small differences between the spectra of rocks. At the same time, the intensity and waveform characteristics of spectral band are fully considered, and the data noise is effectively reduced, compared with the traditional lithology classification method, the hyperspectral angle mapping method (SAM). The improved algorithm can distinguish the similar spectrum of rock more accurately, identify the lithology easily, and have better ability to distinguish the rock with the phenomenon of "same spectrum of foreign body". The SAM method is applied to the lithology classification of TASI data in Liuyuan area of Gansu Province. It can be seen that SDEM method can recognize the lithology of SAM which is not recognized or wrong. The results of lithologic classification obtained by SDEM method are more in line with the actual distribution of rocks, and visible spectrum discrete energy level waveform matching method has better lithologic classification effect and can distinguish ground objects better.
【作者单位】： 中国地质大学(北京)地球科学与资源学院;
【基金】：中国地质调查局项目“热红外高光谱矿化蚀变矿物提取方法研究与应用示范”(12120113099600)
【分类号】：P627
【正文快照】： 0引言高光谱遥感具有多而窄的连续光谱,光谱分辨率可达纳米级,可获得影像中每个像元的精细光谱,包含了表现地物空间分布特征的空间、辐射和光谱三重信息,具有“图谱合一”的特点[1]。地物的光谱特性是高光谱遥感的基础。岩石热红外波段发射率光谱特征与其晶格中原子基团的基频，

本文编号：1370931