Research on Brand Traceability of Mask Physical Evidence Based on Micro-Raman Spectroscopy and PCA-LDA/QDA
A brand traceability analysis model of mask physical evidence was constructed.35 samples of blue disposable surgical masks were tested by micro-Raman spectroscopy,a total of 105 Raman spectral data were collected,and the collected data were pre-processed by Principal Component Analysis(PCA).Combined with Linear Discriminant Analysis(LDA)and Quadratic Discriminant Analysis(QDA),PCA-LDA and PCA-QDA brand traceability analysis models were constructed.The results of micro-Raman spectroscopy showed that all 35 mask samples were made of polypropylene fiber.The PCA-LDA model a-chieved 91.67％and 90.48％traceability accuracy in training set and test set,showing high accuracy and reliability.In contrast,although PCA-QDA model could be applied to mask brand traceability,it had the problem of overfitting.The research showed that the model based on PCA-LDA had better perform-ance and could quickly,accurately and reliably realize the traceability of mask brands,which provided an effective technical means for the forensic scientific analysis of the mask evidence and helped to im-prove the efficiency of the detection of criminal cases.
NETL
NSTL
万方数据
