| This paper consists of the following researches;1. Horizontal attenuated total reflectance Fourier transform infrared spectroscopy (HATR-FTIR) techniques were used to obtain the FTIR spectra of flower of Sophorajaponica L., flower of Dolichos lablab L., and flower of Jasminum sambac (Linn.) Aiton. Then, Fourier self-deconvolution technique is used to transform their FTIR spectra. The results showed that this method could enhance the resolution of small bands that have been buried under other bands in the original spectra and be used to recognize different plants of different genuses and families simply, rapidly, and accurately. Then, Fourier self-deconvolution technique is sequentially used to transform the spectra of flower of Sophora japonica L. and its confusable varieties of flower of Sophora japonica f. pendula Hort. Because they belong to sibling plants and they contain similar chemical composition, and their IR absorption values at the same wave number is quite similar. It is difficult to distinguish between them. In order to make the information that buried in the IR overlapping bands and the difference that existed in the infrared absorptions spectra be displayed visually for the identification of the spectra which are similar, discrete wavelet transform (DWT) was used to extract their features for further classification. This technique eliminates the noise and enhances the signal and the resolution. The results showed that this method can be used to recognize plants of the same genuses and families whose chemical composition and IR absorption are very similar simply, rapidly, and accurately. We successfully established the method of Infrared Spectroscopy based on discrete wavelet transform application to Chinese Medicines for the first time. On this basis, the method of infrared spectroscopy based on discrete wavelet transform application to identify flower of Sophorajaponica L. and its confused species flower of Robinia pseucdoacaci.; flower bud of Sophorajaponica L. and its confusable varieties of flower bud of Robinia pseucdoacaci., Chrysanthemum indicum L., Humulus scandens (Lour.)Merr., Mahonia fortunei (Lindl.)Fedde; matured seed of Prunus persica (L.) Batsch, Prunus armeniaca L.and their confusable varieties.The results showed that this method is universal and practical and could be used to recognize Chinese medicines simply, rapidly, and accurately. This method further improves the accuracy and selectivity of Chinese medicines quality evaluation method, and will play a positive role in the standardization of Chinese Medicines quality and in the future will become one of the ideal methods in the Chinese Medicines quality evaluation system.2. This paper systematically studied the combination of discrete wavelet transform and methods of clustering analysis (CA) and principal component analysis (PCA) that were used to identify common Chinese medicines matured seed of Prunus persica (L.) Batsch, Prunus armeniaca L.and their confusable varieties of grafting almond from different places. The analysis results are consistent with the traditional identification methods using for their quality evaluation. This species related analysis method is intuitive and scientific and has a good application value in the classification and identification of traditional Chinese Medicines.3. We extend the study of the method of Infrared Spectroscopy based on discrete wavelet transform, combining with probabilistic neural network (PNN) application to the identification of Chinese Medicines Bidens frondosa L. and Bidens bipinnata L. The study provides a scientific research foundation for spectroscopy combining with computer science and technology broader application to the identification of Chinese Medicines. |
PDF Full Text Request