拉曼光谱结合聚类分析法区分检验塑钢窗

发布时间：2018-04-25 12:34

  本文选题：塑钢窗 + 拉曼光谱　； 参考：《化学研究与应用》2017年09期

【摘要】：应用拉曼光谱法并结合聚类分析对26种不同品牌、厂家、型号的塑钢窗进行了深入的检验研究。采用显微激光拉曼光谱分析技术分别对样品进行检测,得到在780nm激光光源的一阶导数扩展拉曼光谱中,塑钢窗的光谱形态差异显著,荧光背景干扰弱,重叠的谱峰得到有效的分离,可构建具有高鉴别能力的聚类分析模型,采集光谱数据将其定量化,选择离差平方和法作为类间距离,采用欧氏距离作为度量区间表征样品之间的亲疏程度,进行系统聚类分析,同时结合多种方法验证衡量聚类效果,成功将26种塑钢窗样本分为了18类,实现了基于全波段光谱信息结合系统聚类分析建立模型用于准确鉴定塑钢窗种类的目的,为现场物证的区分检验方法提供了一定的理论依据。
[Abstract]:Raman spectroscopy and cluster analysis were used to examine 26 different brands, manufacturers and models of plastic and steel windows. In the first derivative extended Raman spectroscopy of the 780nm laser source, the spectral morphology of the plastic window was significantly different, and the fluorescence background interference was weak. The overlapping spectral peaks can be separated effectively, and a cluster analysis model with high discriminant ability can be constructed. The spectral data are collected and quantified, and the deviation square sum method is chosen as the inter-class distance. Euclidean distance is used to measure the degree of familiarity between samples, and the cluster analysis is carried out systematically. At the same time, 26 kinds of plastic window samples are divided into 18 categories successfully. The purpose of establishing a model based on the full band spectral information and cluster analysis system to accurately identify the type of plastic and steel window is realized, which provides a certain theoretical basis for the method of distinguishing and checking the material evidence in the field.
【作者单位】： 中国人民公安大学刑事科学技术学院;云南大学信息学院;四川警察学院刑事科学技术系;
【分类号】：D918.9;O657.37
，

本文编号：1801354