摘要
在交通肇事逃逸和刑事案件中,常涉及对汽车前保险杠物证的检验鉴定,本实验采集了31款车型的汽车前保险杠碎片的红外光谱图,借助分类算法,结合多元统计学中的主成分分析法和判别分析法,建立数学模型,对光谱数据逐步展开了挖掘与分析,成功将31种汽车前保险杠样本分为了17类,分类正确率达92.5%,分类结果理想。本研究将现代仪器分析技术和多元统计法相结合实现了对汽车前保险杠碎片的分类和识别,为现场物证的检验提供了一定的理论依据。
It often involves examination and identification about front bumpers in the traffic casualties and escape and criminal cases.This paper was combined the principal component analysis(PCA)and discriminant analysis with Infrared spectra data of 31 front bumpers debris and has established the mathematical model via Classification algorithm.Finally,31 front bumpers debris were successively distinguished into 17 classes and the result was ideal.In this paper,the modern instrument analysis technology and multivariate statistical method were combined to realize the classification and identification of front bumpers debris,which provided a scientific basis for the inspection of physical evidence.
出处
《光散射学报》
2018年第1期70-76,共7页
The Journal of Light Scattering
关键词
汽车前保险杠!红外光谱
判别分析
种类鉴别
front bumpers
infrared spectroscopy
discriminant analysis
identification