摘要
为建立一种简单、快速、无损的面巾纸检验分析方法,利用红外光谱和拉曼光谱对60种不同厂家、不同品牌的面巾纸样品进行检验。采用主成分分析法提取样品光谱数据主成分,用K均值聚类法将样品的两种光谱数据同时分为6类,应用Fisher线性判别和非线性的多层感知器构建判别模型,并对两个模型进行比较。结果表明,Fisher线性判别相对稳定,多层感知器优化后的非线性判别模型可将60种样品分为8类,判别效果较好。结合样品生产原料,对样品分类判别准确率达到93.8%,能较好地对不同产地、不同原料的面巾纸样本进行分类,该方法可为检验面巾纸类物证提供技术参考。
In order to establish a simple,rapid,and nondestructive analytical method for the classification of facial tissues,a total of 60 facial tissue samples from different manufacturers and brands were analyzed using infrared spectroscopy and Raman spectroscopy.Principal Component Analysis(PCA)was employed to extract key features from the spectral data,and K-means clustering was used to classify the samples into six groups based on both sets of spectral information.Subsequently,two discrimination models were constructed:a Fisher Linear Discriminant Analysis(FLDA)model and a nonlinear multilayer perceptron(MLP)model.The performance of both models was compared.While FLDA demonstrated stable results,the MLP-based nonlinear model achieved superior classification performance,grouping the 60 samples into eight distinct categories with a high degree of accuracy.When combined with information on raw material composition,the overall classification and discrimination accuracy reached 93.8%,effectively distinguishing facial tissue samples from different sources and material bases.The established classification method provides a valuable technical reference for the forensic examination of facial tissueevidence.
作者
姜红
黄艺驰
梁爽
陈越
JIANG Hong;HUANG Yichi;LIANG Shuan;CHEN Yue(Department of Criminal Science and Technology,Hunan Police College,Hunan Changsha 410138;College of Investigation,People's Public Security University of China,Beijing 100038;Center of Forensic Science Beijing Hui Zheng Zhuo Yue Technology Co.,Ltd.,Beijing102446)
出处
《中国刑警学院学报》
2025年第3期5-15,共11页
Journal of Criminal Investigation Police University of China
基金
2022年度食品药品安全防范山西省重点实验室开放课题资助项目(编号:2022040709510106)。
关键词
红外光谱
拉曼光谱
面巾纸
FISHER判别
多层感知器
infrared spectroscopy
Raman spectroscopy
facial tissues
Fisherlinear discriminant analysis
multilayerperceptron
