Study On Quantitative Feature-analysis And Recognization Of Radix Ophiopogonis,Pseudostellariae And Polygonat?m

Objective(1)To measure the appearance characteristics and the microscopic characteristics of radix ophiopogonis,polygonat?m and radix pseudostellariae from different parts by combining the Image Processing and camera technology and then to recognized the pattern of radix ophiopogonis,Polygonat?m and radix pseudostellariae by using BP neural network,to achieve an objective automatically as far as possible by using the Image Processing Software to write the suitable algorithms.(2)To determine the saponins' absorbance from different habits and the histochemistry study,and to establish the correlation between microscopic characteristics and the chemical contents,to provide a new idea and method for the identification and the quality evaluate of Traditional Chinese Herbal Medicines.Methods:Firstly,we selected the required number of samples from the overall,put the samples on the white paper and took their macro photos.Secondly,we made transverse microscopic model with PEG 2000 as the embedding media then used the ToupView software and RuihogeRR300C to take the photos of each view while kept some repeated area between adjacent views.Subsequently,we processed the pictures via the Matlab 2014a and the Image pro plus 6.0 and measured the appearance characteristics,microscopic characteristics and chemical contents of three Traditional Chinese Herbal Medicines.Finally,we found the main distinction of different habitats and different positions and established the Eigenvector of different herbals.According to the eigenvector,we built the neural network identification system of different Traditional Chinese Herbal Medicines and quantitatively analyzed the correlations between characteristic parameters and chemical contents.Results:(1).The form color,size,and surface texture features of radix ophiopogonis polygonat?m and radix pseudostellariae could be described quantitatively.(2).The way that can mosaic a lot of microscopic images could be constructed.(3).The neural network model which used to recognize the radix ophiopogonis,polygonatum and radix pseudostellariae was established.(4).The correlation between the feature parameters and the chemical compositions was observed in the study.(5).The features of radix ophiopogonis could be automatically extracted using the technology.Conclusions:1.We can describe the characteristics of samples quantitatively by combining the Image Processing and camera technology,and build the neural network model that can be used to identity Traditional Chinese Herbal Medicines form different habitats.2.The study can describe the features of the radix ophiopogonis,polygonatum and radix pseudostellariae more accurately and objectively.3.We can lay a foundation for further study of identification of Traditional Chinese Herbal Medicines,meanwhile,the correlation between feature parameters and the chemical compositions can reveal the relationships between traditional recognized features and the chemical composition obviously.4.Our study can provide the theoretic fundament of morphologically for appraises and controls the quality of Traditional Chinese Herbal Medicines.