差分拉曼光谱结合聚类分析检验电线塑料外皮研究

Study on Differential Raman Spectroscopy Combined with Cluster Analysis for Inspection of Plastic Wire Sheath

为建立一种高效准确的差分拉曼光谱检验电线塑料外皮的方法,利用便携式差分拉曼光谱仪测得了35个不同品牌不同类型的电线塑料外皮的差分拉曼光谱数据;再根据样品的主要成分和所含填料的不同,对样品进行初步分类;利用主成分分析对初步分类后的拉曼数据进行降维,再利用系统聚类将样品分组,最后应用K-均值聚类分析对分组结果准确性进行检验。其中,对于同组样品,可以用拉曼特征峰的相对峰高比进行区分。结果表明,当并类距离为1时,I-1组样品(只含滑石粉的白色样品)可分为7个小组,实现了对样品的区分;该方法不破坏检材,操作简单,可为微量物证检验和公安机关办案提供帮助。

To establish an efficient and accurate differential Raman spectroscopy method for the detection of wire plastic skin,the differential Raman spectroscopy data with 35 different brands and types of wire plastic skin were measured by a portable standard differential Raman spectrometer. The samples were preliminarily classified according to their different main components and fillers. A principal component analysis was used to reduce the dimension of the Raman data after preliminary classification. The samples could be grouped by hierarchical clustering,and a K-means clustering analysis was used to test the accuracy of the grouped results. For the same group of the samples,the relative peak height ratio of Raman characteristic peaks was used to distinguish. The results indicated that the samples of group I-1(white sample containing only talc)were divided into 7 groups at a distance of 1,and they could be distinguished. This method is easy to operate without any damages in the sample. It can provide help for the detection of trace evidence and the public security organs to handle cases.