基于拉曼光谱的混合矿物颜料检测分析

发布时间：2018-02-25 19:37

本文关键词： 拉曼光谱定量分析解析模型矿物颜料　出处：《长春理工大学》2017年硕士论文　论文类型：学位论文

【摘要】：拉曼光谱检测技术以其高灵敏度、所需检测样品量少、无需损坏样品和原位检测等优点,在对古代艺术品(如壁画、绘画、雕塑等)的检测、矿物颜料的分析及原位修复过程中起到日益重要的作用。本文主要完成的研究内容如下:1.在532 nm和785 nm两种激发光源下,对一些在古代壁画艺术品中常用的红色系、黄色系、白色系等矿物颜料进行拉曼光谱的检测。实验结果显示,在国内外相关研究较少的低波数(40~200 cm-1)范围内,我们选取的大部分样品均有明显拉曼特征峰,而这一部分数据对古代壁画等艺术品的分析、真伪鉴别、原位修复等均有着重要的参考价值。2.在对混合矿物颜料不同成分的定量分析过程中,将模式匹配算法引入到拉曼光谱分析的应用中,测量了古代艺术品常用的雌黄和蛤白混合矿物颜料的拉曼光谱,并对光谱进行了解析,对光谱进行了相似性匹配实验分析。再将变量组合集群分析法与偏最小二乘法结合建立VCPA-PLS模型引入对混合矿物颜料的定量分析中,并对两种算法的结果进行比较,取得了可信度较高的实验结果。3.将BOSS模型引入到对较为复杂的三种混合矿物颜料进行定量分析,该方法的优势在于即使建模数据和预测数据相对较少时,也能够达较高预测精度,其预测数值高于VCPA-PLS分析模型。
[Abstract]:Because of its high sensitivity, small amount of samples, no need to damage samples and in situ detection, Raman spectroscopy is used to detect ancient works of art (such as murals, paintings, sculptures, etc.). The analysis and in situ restoration of mineral pigments play an increasingly important role. The main contents of this paper are as follows: 1. Under two excitation light sources of 532 nm and 785 nm, some common red and yellow systems in ancient mural art are studied. The Raman spectra of white pigments and other mineral pigments were tested. The experimental results show that in the range of 40200cm-1 of low wave number, which is less studied at home and abroad, most of the samples we selected have obvious Raman characteristic peaks. This part of the data has important reference value for the analysis of ancient mural paintings and other works of art, identification of authenticity, in situ restoration, etc. 2. In the process of quantitative analysis of different components of mixed mineral pigments, The pattern matching algorithm is introduced into the application of Raman spectrum analysis. The Raman spectra of the mixed mineral pigments of female and clam white, which are commonly used in ancient art, are measured, and the spectra are analyzed. The similarity matching experiment was carried out on the spectrum. Then the VCPA-PLS model was established by combining the variable combination cluster analysis method with the partial least square method, and the results of the two algorithms were compared with each other in the quantitative analysis of mixed mineral pigments. The experimental results with high reliability are obtained. The BOSS model is introduced to the quantitative analysis of three mixed mineral pigments. The advantage of this method is that even when the modeling data and prediction data are relatively small, The predicted value is higher than that of VCPA-PLS model.
【学位授予单位】：长春理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O657.37;TQ620.72

【相似文献】