摘要
建立便签纸物证快速、准确的识别分类模型,为基层公安获取线索提供数据和理论依据。利用显微共聚焦激光拉曼光谱法(LCM-Raman)检验44个便签纸样品,结合主成分分析(PCA)、系统聚类(HCA)和线性判别分析(LDA)实现便签纸样品识别分类。依据样品中无机填料的拉曼光谱特征峰的不同,对样品进行定性分析并初步分为19类。为了更加科学准确地对样品进行分类,利用主成分分析对样品拉曼光谱数据降维,并利用系统聚类结合线性判别分析确定样品最佳分类为4类。结果表明,LCM-Raman结合PCA-HCA-LDA分类模型可区分便签纸样品,分类结果科学准确。
To establish a rapid and precise identification and classification model for material evidence on note paper,44note paper samples underwent Laser Confocal Micro-Raman(LCM-Raman)analysis.By integrating Principal Component Analysis(PCA),Hierarchical Cluster Analysis(HCA),and Linear Discriminant Analysis(LDA),the samples were identified and classified.Firstly,according to the different Raman spectrum characteristic peaks of inorganic fillers in the samples,the samples were qualitatively analyzed and preliminarily divided into 19categories.Secondly,for the sake of a more scientific and accurate classification,The principal component analysis was used to reduce the dimension of the Raman spectrum data of the samples,and the system clustering and linear discriminant analysis were used to determine that the best classification of the samples was 4categories.The samples from the same place of origin had good aggregation,and the classification accuracy and cross validation classification accuracy were both 100%.Finally,The PCA-HCA-LDA classification model was used to predict the ungrouped sample of note paper.As concluded from the results,LCM-Raman combined with PCA-HCA-LDA classification model can distinguish note paper samples The generalization accuracy of the model is 100%,and the classification results are scientific and accurate.
作者
陈壮
贾成贺
姜红
Chen Zhuang;Jia Chenghe;Jiang Hong(Judicial Police Academy(Public security branch),Gansu University of Political Science and Law,Lanzhou730070,Gansu,China;Institute of Evidence Law and Forensic Science,China University of Political Science and Law,Beijing 100080,China;Public Security Department of Xinjiang Uygur Autonomous Region,Urumq 830000,Xinjiang,China;Department of Criminal Investigation,Gansu Police Vocational College,Lanzhou730046,Gansu,China)
出处
《应用激光》
CSCD
北大核心
2024年第2期94-103,共10页
Applied Laser
基金
甘肃省教育厅高校教师创新基金项目(2023A-100)
甘肃政法大学2020年“引才专项”项目(gszf2020xyc004)
甘肃政法大学2023年度校级科研创新项目(GZF2023XQN08)。
关键词
光谱学
显微共聚焦激光拉曼光谱法
主成分分析
系统聚类
线性判别分析
便签纸
spectroscopy
laser confocal micro-raman
principal component analysis
hierarchical clustering analysis
linear discriminant analysis
note paper
