The Research Of Fluorescence Spectral Imaging Applied To Detect Quality Of Traditional Chinese Medicine

With the advantages of celerity, nondestructive, low cost and greenery, Fluorescence spectral imaging has become an important way in the field of detection and analysis. Fluorescence spectral imaging has highly academic value and application value in detecting the quality of traditional Chinese medicines(TCM). We can use Fluorescence spectral imaging to research on quality detection of TCM accurately, directly, rapidly, so that the potency of TCM can be more effective. In this paper, we use Fluorescence spectral imaging technology combined with chemometrics and digital image processing to analyze the TCM directly and quickly.In the first part of this paper, we identify the species of the Cistanche by using Fluorescence spectral imaging technology combined with pattern recognition methods. Firstly, we use principal component analysis(PCA) and Fisher discriminant analysis to identify the species of the Cistanche. After we mix two of different kinds of Cistanche, and then use gray threshold segmentation method to distinguish the mixed powder of Cistanche by color RGB image. In the second part of this paper, we study the active ingredient of pilose antler pieces by Fluorescence spectral imaging technology. We draw the color map in both 2D and 3D visualization after extract the feature dimension band of the spectrum cube of pilose antler pieces. The results combined with biological medicine identification show that the active ingredient of pilose antler pieces are different due to the varieties of species, medicinal parts and biological tissue. In the third part of this paper, we apply Fluorescence spectral imaging technology, to study on the influence of the size of TCM powders on mixed uniformity. We select two of the medicines in Angong Niuhuang Wan--- Coptis and Curcumae, and mix them together. Through moving block macropixel relative dtandard deviation(MBMSTDEV), the uniformity of the blend powders can be evaluated. The main achievements of this paper:(1) We can identify the Cistanche through Fluorescence spectral imaging technology combined with pattern recognition. Firstly, we can get the first four main components of the Cistanche by principal component analysis. Then we use the data of three principal components, which cumulative contribution rate are over 80%, to draw scatter plots. After that, we can identify the different types of Cistanche generally. Furthermore, the following results can be achieved after establish the model by fisher discriminant analysis: the discrimination rate of initial case is 100%, the avvuracy of cross case discriminant is 95%. We choose any two different types of Cistanche powder that their volumes are equal, and then mix them together. The mixed powder can be classified through thresholding segmentation method. Also, the mixed powder color map in red and green can be achieved For an example, the mixed powder color map of Cistanche tubulosa and Cistanche deserticola show that the proportion of green part and red part accounts for1:1.12.(2) The active ingredient of pilose antler has been studied through applying Fluorescence spectral imaging technology. Take the dry antler form Cervus Nippon Temminck(A8), dry antler form Cervus elaphus Linnaeus(A10) and sand antler form Cervus Nippon Temminck(A3) as examples, we study the different active ingredient in pilose antler as varied as their species, medicinal parts or biological tissue. First of all, we can get the following results after compare the characteristic from spectral curve: ①The active component of A8 is higher than A10. ②A10 contains more active ingredients than A3. ③ According to the different parts of the organization, medulla layer of the pilose antler has highest active ingredientbut stromal layer has least. We can get the following results aftet compared the characteristic from the aspect of space. ①The active ingredient of A8 and A10 are concentrated in medullary layer. ②The active ingredient of A3 are low and distribute dispersedly. ③The active ingredient of pilose antler are very different as their different biological tissues. We can not only classify the pilose antler, but also distinguish the three biological tissues basically by PCA and image enhancement methods.(3) We study different size of traditional Chinese medicine power on the impact of the mixing unformity of the solid compound preparation conbine with Fluorescence spectral imaging technology. Firstly, we choose two flavor prescription of Angong Niuhuang Wan--- Coptis and Curcumae to establish binary mixing intermediates. Secondly, we mix four different particle sizes(60 mesh, 80 mesh, 100 mesh, 120 mesh) of Coptis and Curcumae powder. After that, the minimum correlation but maximun amount of information of the two band image in the mixed powder spectral can be gained. Before enhance the image, we calculate the image ratio.And then we use thresholding method to classify the mixed powder of Coptis and Curcumae. Finally, we use phase cell block standard deviation to evaluate the uniformity of different size mixed powder of Coptis and Curcumae. The results show that 80 mesh can be easier to blend uniformly compared to 100 mesh.