Raman spectrum is the fingerprint of the molecule, and Raman peak locations relate to the structure of the molecule, so identification of specimen can be performed based on Raman spectroscopy. However, classification of Raman peaks of the mixture is usually difficult, especially, when Raman peaks of one component are overlapped with, even dominated by those of other component. As a nonlinear dynamic system, artificial neural network simulates the functions and characters of biological neural systems, and has been widely used in signal process. Firstly, this paper introduced the principles of Raman spectroscopy and the technique of artificial neural network. Secondly, Raman spectra of the mixtures of CCl4 and ClO4 - were analyzed using BP artificial neural networks. The results showed that artificial neural network technology is flexible, affordable and easily adapted to qualitative analysis of Raman spectrum of the mixture.

Date Published: 10 July 2009
PDF: 5 pages
Proc. SPIE 7490, PIAGENG 2009: Intelligent Information, Control, and Communication Technology for Agricultural Engineering, 749029 (10 July 2009); doi: 10.1117/12.836648

Published in SPIE Proceedings Vol. 7490:
PIAGENG 2009: Intelligent Information, Control, and Communication Technology for Agricultural Engineering
Honghua Tan; Qi Luo, Editor(s)