| X-ray diffraction(XRD) having been an important method involved in modern traditional Chinese medicine analysis, some problems yet exist for Tibetan medicine, one of which lies in the data processing of XRD patterns. The fingerprints of most of Tibetan medicine, especially the plant-based, exhibit a vague disperse broad peak with massive ‘burr’ and fairly few characteristic diffraction peaks overlapping, owing to the medicine containing some macro-molecular organic substances such as starch, saccharose, protein,and so on, belonging to elementary syngony, which leads to fairly difficult patterns analysis and restrain its application on identification and quality control of Tibetan medicine. Some researches conducted in this paper are just aimed at resolving this problem, and the detailed contents include four main parts as follows. For the first part, based on professional XRD pattern analysis software such as MDI Jade5.0 and X’Pert Highscore Plus, the implementation in Jade5.0 and relevant algorithms of traditional data pre-processing modules such as format transformation, pattern smoothing, background&K? removal and peak searching, were introduced. A comparison study on denoising effects of XRD patterns was conducted between Savitzky-Golay filter, wavelet Modulus Maximum denoising algorithm and multi-interval threshold denoising algorithm.For the second part, a procedure utilized for obtaining powder X-ray diffraction files was put forward to resolve the problem of difficult pattern analyzing of plant-based Tibetan medicine, which encompasses four main steps, that is, gaining XRD pattern files, gaining second derivative pattern files, peak searching and making powder X-ray diffraction files, in which Savitzky-Golay filter and symmetric zero-area convolution peak searching algorithm were used. In this part, we made a brief introduction in principle for second derivative patterns calculating of discrete data such as three-points formula method, five-points formula method, cubic spline method and cubic B-spline method. Based on MALAB, we developed a set of programs, which can obtain relevant information of Tibetan medicine powder X-ray diffraction files, conveniently and fast.For the third part, the X-ray Diffraction Whole Pattern Peak-resolution Software of Plant Tibetan Medicine was designed, which includes five subsystem modules of format transformation, pattern smoothing, background removal, initial peak searching and peak-resolution fitting, covering wavelet Modulus Maximum denoising algorithm, sensitive nonlinear iterative clipping algorithm and least square fitting. Based MATLAB and the above algorithms, the software for making XRD pattern peak-resolution was developed, which possesses interactive graphic interface and is quite easy for user to do peak-resolution work and obtain abundant diffraction peaks information consisting of peak locations, peak intensity, full width at half maximum(FWHM), peak-shape factor and skewness factor, providing reference for further medcine analysis such as crystallinity calculation and similarity calculation. For the forth part, based on vectorial angle method, correlation coefficient method and relative entropy theory, a similarity calculation procedure was given for XRD patterns, including four steps of original data collection, normalization, feature information extraction and similarity calculation.The mentioned above are just the main research contents and consequences. In general, three methods, that is, powder X-ray diffraction files gaining procedure, X-ray Diffraction Whole Pattern Peak-resolution Software of Plant Tibetan Medicine and pattern similarity calculation mode were studied and designed for the XRD pattern analyzing of Tibetan medicine, especially the plant-based. It is believed that the research conducted in this paper will play an active role in performing efficient fast accurate identification and quality control of Tibetan medicine and the XRD patterns’ digitalization of Tibetan medicine. |
PDF Full Text Request