Rapid Discrimination of Jade Origins Based on Molecular Spectra

The spectral characteristics of jades from Burma, Russia, and Guatemala are compared and analyzed. Based on the infrared spectroscopy and Raman spectroscopy combined with the principal component analysis (PCA) and the back propagation (BP) neural network, the model is built for discriminating jade origins, whose discrimination effect is tested. The results show that the infrared absorption spectra of jades from different origins are basically similar, but there exists certain difference in their Raman spectra. In addition, Jades from Burma, Guatemala, and Russia all possess the characteristic Raman spectra. Based on the PCA-BP neural network discrimination model of infrared spectra, the discrimination accuracies of training samples and test samples are 94.2% and 91.6%, respectively. In contrast, based on the PCA-BP neural network discrimination model of Raman spectra, the discrimination accuracies are 93.48% and 100.0%, respectively. It can be seen that the PCA-BP neural network discrimination models based on infrared spectra and Raman spectra both have high accuracy in identifying jade origins, indicating they have certain practicability and feasibility for rapid identification of jade origins.