Spectral Classification and Identification of Methamphetamine and Its Common Additives Based on Multivariate Modeling

Hou Wei; Wang Jifen; He Xinlong

doi:10.3788/LOP202158.0330003

Laser & Optoelectronics Progress, Volume. 58, Issue 3, 3300031(2021)

Spectral Classification and Identification of Methamphetamine and Its Common Additives Based on Multivariate Modeling

Hou Wei... Wang Jifen* and He Xinlong |Show fewer author(s)

Author Affiliations

School of Investigation, People''s Public Security University of China, Beijing 100038, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(7)

Fig. 1. Infrared spectra of all samples

Download full size

Fig. 2. Classification accuracy of full band spectral data at characteristic variable with different dimensions

Download full size

Fig. 3. Classification accuracy of infrared fingerprint region data at characteristic variable with different dimensions

Download full size

Fig. 4. Spatial classification details of five additives

Download full size

Table 1. Two constituents samples preparation methods
View table
Table 1. Two constituents samples preparation methods
No. Methamphetamine /mg Additives /mg Mass fraction of methamphetamine /%
1 0.5 4.5 10
2 1.0 4.0 20
3 1.5 3.5 30
4 2.0 3.0 40
5 2.5 2.5 50
6 3.0 2.0 60
7 3.5 1.5 70
8 4.0 1.0 80
9 4.5 0.5 90

Table 2. Classification accuracy of five additives under different methods
View table
Table 2. Classification accuracy of five additives under different methods
Type MLP RBF BDA
Training set Test set Training set Test set Training set Test set
Fingerprint spectrum 99.25 94.80 74.55 62.38 98.70 96.67
Full spectrum 98.50 90.12 70.16 60.53 98.70 96.67

Table 3. Classification accuracy of samples containing different mass fractions of methamphetamine under different methods

View table

Table 3. Classification accuracy of samples containing different mass fractions of methamphetamine under different methods

Model	Full spectrum		Fingerprint spectrum
Model	Original variable	Characteristic variable	Original variable	Characteristic variable
MLP	42.67	5.23‒46.58	48.53	6.52‒50
RBF	7.05	0‒23.22	4.31	0‒25.40
BDA	82.30	20.85‒100	82.30	22.78‒100

Tools

Get Citation

Copy Citation Text

Hou Wei, Wang Jifen, He Xinlong. Spectral Classification and Identification of Methamphetamine and Its Common Additives Based on Multivariate Modeling[J]. Laser & Optoelectronics Progress, 2021, 58(3): 3300031

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites