光散射学报, 2018, 30 (2): 168, 网络出版: 2018-08-04   

拉曼光谱结合系统聚类法检验汽车前保险杠

Examination the Front Bumpers Based on Raman Spectroscopy and Hierarchical Clustering
作者单位
1 中国人民公安大学刑事科学技术学院,北京,100038
2 中国人民公安大学侦查与反恐怖学院,北京,100038
3 江西财经大学会计学院,江西 南昌,330103
摘要
应用拉曼光谱法并结合聚类分析对44种不同型号的汽车前保险杠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532 nm激光光源的拉曼光谱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,选择平均组间距离定义类间距离,采用平方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,通过验证衡量聚类效果,成功将44种汽车前保险杠样本分为了19类,分类正确率达94.8%,分类结果理想,实现了基于全波段光谱信息结合系统聚类分析建立的模型用于准确鉴定汽车前保险杠种类的目的,为现场物证的检验提供了一定的理论依据。
Abstract
Study on the species identification of 44 different brands front bumper with Raman spectroscopy and hierarchical clustering.Using micro laser Raman spectroscopy technology detects samples respectively,and showed that in the Raman spectrum first derivative of 532 nm laser propagation,the front bumper spectral morphology could construct a model of clustering with the high discrimination.Gathered the spectral data and quantified it.The characteristic data were clustered by selecting the between-groups linkage as the distance between classes and using squared euclidean as the closeness between interval representations of samples.44 kinds of front bumper were successfully divided into the following 19 categories by measuring the clustering effects and the cross validation of the classification accuracy was 948%,the result was ideal.The model based on the full band spectral information and the Hierarchical cluster could be used to accurately identify the types of front bumper,which provided a theoretical basis for examining the physical evidence.

何欣龙, 王继芬, 刘腾飞, 祁永睿. 拉曼光谱结合系统聚类法检验汽车前保险杠[J]. 光散射学报, 2018, 30(2): 168. HE Xinlong, WANG Jifen, LIU Tengfei, QI Yongrui. Examination the Front Bumpers Based on Raman Spectroscopy and Hierarchical Clustering[J]. The Journal of Light Scattering, 2018, 30(2): 168.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!