| Quality control and quality evaluation of traditional Chinese medicine(TCM)is one of the key problems that restrict the development of the modernization of TCM,and it has been a difficult and hot spot in the research of traditional Chinese medicine.The current Chinese medicine quality control mode is established according to the foreign plant medicine and chemical medicine.However,due to the complexity ofTCM,it is hard to effectively evaluate the quality and to reflect its safety and effectiveness.For this reson,this paper proposed the theory of quality evaluation of TCM based on surface color grading.Moreover,coloration is a serious phenomenon in market.In this paper,safflower was taken as an example to study the color grading method and coloration problem.Due to the subjectivity and fuzziness of the visual method,the color can only be roughly estimated and brief descriptedusing language.In this paper,the objective characterization and accurate quantification method of TCM color was established using machine vision technology,through the image acquisition,image segmentation,color space conversion and color feature extraction.The optimal condition of image acquisition was optimized by using the colorimeter,and the results showed that the estimated color were reliable under the optimal condition.The collected 30 batches of safflower pieces were analyzed and classified using the conventional analysis method,and the HPLC fingerprint of safflower was established.The results showed 9 batches were stained.According to the hydroxysafflor yellow A(HSYA)content,water leaching content and similarity of fingerprint chromatogram,21 batches of unstained safflower wasclassfied into 3 quality levels.High quality safflower:HSYA content is 1.32%-1.64%,water leaching content is 38.21%-41.86%.The L value is in the range of 37.01-42.14,a value is 37.67-41.34,b value is 34.60-41.40.Average quality safflower:the concentration of HSYA is 1.04%-1.45%and water leaching content is 32.28%-40.40%.L value range is 36.34-40.56,a value is 32.64-38.45,b value is 33.21-38.89.Unqualified safflower:the content of HSYA is 0.14%,water leachingcontent is 10.99%.L value is 35.03-35.91,a value is 30.13-30.79,b value is 29.01-30.11.The relationship between HSYA,water leaching content and color values was analyzed using correlation analysis.Results showed effective component content and color values are closely related.HSYA and water leaching contents decreased gradually when surface color fade and become darker.Multivariate linear regression equationswere established to predict the component content according to color values.The HSYA content equation is Y=73.964S+3.047V-0.228L-0.095C+0.450B-59.878(R2=0.709).The water leaching content equation is Y=1408S+178.167V-6.859L-1.669C+8.121B-1095.2(R2=0.806).Results showed that it is feasible to predict the content of chemical components by appearance color.Partial least squares(PLS)quantitative models of HSYA and water leaching content were established based on the color values and near infrared spectroscopy.Integrated quantitative models were built by using data fusion method.The results showed that the integrated model has better performance than the single model.The prediction results have a great improvement,especially for HSYA,which has low content.The distribution differencesof color L,a and b values between safflower and dyed safflower were studied.The L value distribution of undyed safflower is 35-43,a is 30-42 and b is 28-44,which are concentrated than dyed safflower.Then,the L value,a value,b value and C value were combined to identify the dyed flowers,and the success rate was 88.89%.The raw NIR spectra,second derivative spectra and two-dimensional correlation spectroscopy of six commonly used staining agents were studied for identification,which were lemon yellow,auramine O,sunset yellow,orange II,neucoccin and acid red 73.The results showed that there was a significant difference between the colored safflower and the non-stained safflower.Different dyeing agents have different absorption peaks,which could be used for identification.PLS model ofsix different dyeing agents is established.The results showed PLS model established with lemon yellow characteristic band 4200-4600 cm-1,auramine O band 5991-7162 cm-1,sunset yellow band 4050-4700 cm-1,orange ? band 4050-4700 cm-1,neucoccin band 4125-4456cm-1,acid red 73 band 5000-5330cm-1has better performance than raw spectra.It was demonstrated the selected characteristic bands can be used for modeling ofdifferent dyeing agents.In summary,this paper proposed a novel method for quality evaluation of TCM based on color values.A quantitative characterization method was established to determine the color of TCM,which provide modern technical support for traditional identification method.The relationship between color and component content was analyzed,and the quantitative prediction model was established.The identification and quantitative analysis of different dyeing agents were realized by near infrared spectroscopy. |
PDF Full Text Request