基于Hapke模型的岩矿反射率光谱模拟研究

发布时间：2018-06-02 04:00

本文选题：Hapke模型 + 岩矿混合光谱　；参考：《中国地质大学(北京)》2015年硕士论文

【摘要】：高光谱遥感技术日趋成熟使定量化分析地物波谱成为可能,遥感地物光谱模型是定量分析手段之一。本文使用IDL编程语言,开发了基于Hapke岩矿光谱模型的“非线性最小二乘法”反射率光谱反演模拟程序,通过各种实测岩矿粉末光谱与模拟光谱对比,采用相关系数R和均方根误差RMSE进行评价,得到该模型在模拟矿物及混合矿物反射率光谱方面具有很高的准确性和可行性;同时利用该模型尝试性地模拟了甘肃柳园地区花岗岩反射率光谱,依据模拟光谱特征对CASI/SASI高光谱数据开展岩性提取工作,经野外验证提取结果准确度较高。论文研究取得如下成果:1.依据Relab光谱库数据和实验室实测多矿物、多角度、多粒径混合反射率光谱,对Hapke各向异性模型(AMSA)和各向同性模型(IMSA)反演模拟岩矿粉末样品反射率光谱的准确性进行了严谨细致的研究评价,综合结果表明AMSA模型准确性更高且可调试能力更好;且AMSA模型各参数间具有高度耦合性,参数设置越准确则模拟精度越高。2.针对岩矿混合反射率光谱,本文对实验室中多组不同矿物、不同混合比例的混合矿物进行光谱实测,并采用AMSA模型反演模拟其反射率光谱,研究结果表明该模型在反演颗粒状混合矿物光谱方面准确性很高,对片状矿物适用性较差,但可通过调整模型相关参数提高反演精度;同时该模型在岩矿混合光谱研究方面具有很高的实用价值,对于混合矿物光谱“非线性”的特性,其可定量分析其矿物组成与矿物丰度,提升遥感人员对岩矿混合光谱的研究认识。3.以花岗岩为例,依据已有工作区花岗岩矿物组成综合分析结果,使用AMSA模型近似模拟花岗岩反射率“纯光谱”,以模拟波谱结合实验室多种花岗岩类岩石实测波谱,总结花岗岩反射率波谱特征规律。并利用此规律采用matched-filtering高光谱波谱识别技术开展岩性提取工作,经野外资料验证花岗岩提取结果准确度较高,表明本文所用提取流程具有实际可操作性。综上所述,本文所取得的研究成果可提升遥感人员对Hapke岩矿光谱模型的认知程度,并帮助其它地区开展类似高光谱岩性填图工作,具有一定指导意义。
[Abstract]:Hyperspectral remote sensing technology is becoming more and more mature, which makes it possible to quantitatively analyze the spectral spectrum of objects, and the spectral model of remote sensing objects is one of the quantitative analysis methods. In this paper, using IDL programming language, a "nonlinear least square" reflectance spectral inversion simulation program based on Hapke rock and ore spectral model is developed. The correlation coefficient R and root mean square error (RMSE) are used to evaluate the model. The model is accurate and feasible in simulating the reflectance spectrum of minerals and mixed minerals. At the same time, the model was used to simulate the reflectance spectrum of granite in Liuyuan area, Gansu Province. According to the simulated spectral characteristics, lithologic extraction of CASI/SASI hyperspectral data was carried out. The research results are as follows: 1: 1. Based on the Relab spectral database data and laboratory measurements, the mixed reflectance spectra of multi-mineral, multi-angle and multi-particle size were obtained. The accuracy of Hapke anisotropic model and isotropic model for inversion of reflectance spectrum of simulated rock and ore powder samples are studied and evaluated in detail. The comprehensive results show that the accuracy of AMSA model is higher and its debugging ability is better. The parameters of AMSA model are highly coupled, and the more accurate the parameters are, the higher the simulation accuracy is. Aiming at the mixed reflectance spectrum of rock and ore, the spectral measurements of mixed minerals with different minerals and different mixing ratios were carried out in the laboratory, and the reflectivity spectra of the mixed minerals were simulated by AMSA model. The results show that the model has a high accuracy in retrieving the spectra of granular mixed minerals, and is less suitable for flake minerals, but the inversion accuracy can be improved by adjusting the relevant parameters of the model. At the same time, the model has a high practical value in the study of mixed spectrum of rock and ore. For the "nonlinear" characteristics of the spectrum of mixed minerals, it can be used to quantitatively analyze the mineral composition and mineral abundance. Improve remote sensing personnel's understanding of rock and ore mixed spectrum. 3. Taking granite as an example, according to the results of comprehensive analysis of the mineral composition of granites in existing workspaces, the AMSA model is used to approximate simulate the "pure spectrum" of granite reflectance, and the simulated spectrum is combined with the measured spectra of various granitic rocks in laboratory. The characteristics of reflectance spectrum of granite are summarized. The extraction of lithology was carried out by using the matched-filtering hyperspectral spectrum recognition technique. The accuracy of the extraction results of granite was verified by field data, which shows that the extraction process used in this paper is practical and operable. To sum up, the research results obtained in this paper can enhance the cognition of Hapke rock and ore spectral model, and help other regions to carry out similar hyperspectral lithologic mapping, which has certain guiding significance.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P627

【参考文献】