基于同色油墨红外光谱的卷烟包装纸真伪鉴别模型

目的:利用卷烟包装上同色油墨的红外光谱差异快速区分卷烟包装纸真伪。方法:分别收集了云烟、玉溪和红塔山3个品牌的真假卷烟包装纸共116个,使用傅里叶变换红外光谱仪(FTIR)衰减全反射(ATR)采集附件采集样品中红色油墨的红外光谱图。对光谱进行二阶导处理后,以光谱范围为3 200~2 700,1 800~1 580,801~720 cm1,56个真品为模型集合,每品牌各20个赝品验证准确性。分别按照品牌区分和不分品牌区分,使用马氏距离判别建立卷烟包装纸真伪鉴别模型。结果:建立了真假卷烟包装纸定性鉴别模型,对云烟、玉溪和红塔山3个品牌的预测准确率为100%,混合模型对随机样品的预测准确率也达到100%。结论:该组模型将同色油墨作为总体样本,预测准确率可达100%,并能大大提高筛查真假卷烟包装纸的工作效率。

An anti-counterfeit models of cigarette packaging paper based on same color ink by FTIR

Objective: Established a group of models used to quickly distinguish anti-counterfeit cigarette packaging paper. Methods: The model based on the same color oil on packaging paper from 3 different brands: Yunyan, Yuxi, and Hongtashan. The Fourier Transform infrared spectroscopy (FTIR) spectrum were collected from 116 samples totally by Attenuated Total Reflection (ATR) accessary. 56 spectrum were used to build the models and 20 samples of each brand were used to verification. Spectrum were second derivative processed and spectral ranges were selected as 3 200~2 700, 1 800~1 580, and 801~720 cm1. Results: Using Mahalanobis distance discriminant method to establish 4 models (3 brands and 1 multiple), all of the accuracy of models reached to 100%. Conclusion: This is the first time to use the same color oil FTIR spectrum to distinguish real and fake cigarette packaging paper which were highly improve selection efficiency for large number of samples.