| Wuchang Daohuaxiang rice,as a representative commodity of China’s high value-added agricultural products,is famous at home and abroad for its unique quality characteristics and food color,aroma and flavor,and has high economic value.At the same time,counterfeiting driven by economic interests has been banned repeatedly.This commodity refers to Wuyoudao No.4 rice planted within the protection scope of Wuchang geographical indications.According to the field visit and investigation,Some illegal traders replaced Wuyoudao No.4 with other high-yield varieties similar to Wuyoudao No.4.In order to avoid regulatory review,they also adulterated other varieties of rice with Wuyoudao No.4 in a certain proportion,which increased the difficulty of review and undermined the market order.In order to study the adulteration and identification of Wuchang Daohuaxiang,six varieties of rice with similar appearance quality to Wuyoudao No.4 were selected from the test samples,and the adulterated rice samples of different varieties of rice and Wuyoudao No.4 were prepared.The adulteration ratio was initially5%,increased to 50% with a gradient of 5%,and the limit value was set as 100%.A total of 231 samples were tested,and the corresponding research results are as follows:(1)Evaluation and identification of rice quality indexes of different adulterated varieties.The quality indexes of seven varieties of rice were evaluated,and the apparent properties and physical and chemical properties indexes of different adulterated varieties of rice were measured.There were very significant differences in chalkiness and chalky grain rate between Wuyoudao No.4 and other varieties of rice(P < 0.01).In terms of apparent properties,there are certain similarities between different varieties of rice of Wuyoudao No.4,which provides a certain possibility for adulteration.Qualitative discrimination of different adulterated rice varieties based on unsupervised and supervised pattern recognition methods.Using principal component analysis,four principal component factors are extracted and combined with UMAP algorithm to reduce the dimension of the data twice,which can preliminarily classify different adulterated rice varieties.Using hierarchical cluster analysis,when the class spacing is 25,Wuyoudao No.4 can be classified and analyzed with other varieties.Six supervised pattern recognition methods were used to model and k-fold cross validation was used to verify the discrimination accuracy of the model.The discrimination accuracy of support vector machine,naive Bayes and random forest algorithm for different adulterated varieties was 100%.When the adulteration amount is 100%,that is,pure varieties are adulterated,the identification of Wuchang Daohuaxiang rice can be realized based on quality index and chemometrics algorithm.(2)Study on identification of adulterated Wuchang Daohuaxiang rice based on FT-NIRWhen the detected samples are adulterated rice samples in a certain proportion,the actual discrimination is more difficult,and the near-infrared spectroscopy technology is mature in the research of rice identification.In order to realize the qualitative discrimination of different adulterated rice varieties and the quantitative discrimination of different adulterated proportion,the experiment is based on the original near-infrared spectrum of samples collected by FT-NIR,and the characteristic spectrum is extracted and screened by principal component analysis and hierarchical cluster analysis.When establishing the qualitative discrimination model of different adulterated varieties,six pretreatment methods and six supervised pattern recognition methods were investigated.The k-fold cross validation method was used to verify the discrimination rate of the model.Through comprehensive evaluation,it was found that the standard normal variable transformation pretreatment method has good application potential.The qualitative discrimination model based on random forest algorithm has the best effect,and the discrimination accuracy of different adulterated varieties is up to 100%.When establishing the quantitative discrimination model of different adulteration proportion,the discrimination model based on partial least squares discriminant analysis method has a determination coefficient of more than 0.99 and the root mean square error of the prediction set is within the ideal range,indicating that the model has good accuracy and stability.(3)Identification of Wuchang Daohuaxiang rice based on HS-SPME-GC-MSThere are non fragrant rice varieties in the collected samples.There are certain similarities in physical and chemical properties between non fragrant rice and Wuyoudao No.4,and there is a fitting situation in the absorbance data in the specific wavelength range of near-infrared spectrum.Experienced merchants can judge the authenticity of Wuchang rice by smelling the rice flavor,but this method has certain subjectivity and limitations.In order to realize the specific identification of volatile components of Wuchang rice,the fingerprint information of volatile components of different adulterated rice samples was extracted and identified based on HS-SPME-GC-MS.Through the single factor test,the test conditions are gradually optimized,and the best test condition is 50 / 30 μm PDMS-DVB-CAR extraction head,sample volume 15 g,equilibrium temperature 80 ℃,equilibrium time 40 min,extraction time 30 min,resolution temperature 230 ℃,resolution time 1 min.On this basis,127 components with relative volatile content greater than 0.05% were extracted and identified from different adulterated rice varieties.There are 62 kinds of volatile components.Principal component analysis and hierarchical cluster analysis were used to identify different adulterated varieties,and 18 characteristic volatile components were selected from the common volatile components.Establish the standard volatile fingerprint of Wuyoudao No.4,calculate the similarity between the adulterated rice sample and the standard fingerprint based on the characteristic volatile components,and establish the adulteration discrimination model based on the linear fitting equation.The similarity between the adulterated sample of Suijing No.18 and Wuyoudao No.4 is the highest,and the similarity between the adulterated sample of non fragrant rice variety is the lowest,Different adulterated varieties can realize quantitative discrimination within a certain adulterated proportion,and can preliminarily realize the identification of adulterated Wuchang Daohuaxiang rice in theory.(4)Discriminant analysis of adulterated rice samples by feature information fusionIn order to further improve the accuracy of the identification of Wuchang Daohuaxiang rice,reduce the difficulty of test detection and data analysis,and break the limitations of single source data analysis,the experiment used the screened characteristic spectrum and volatile component information to identify different adulterated rice samples.The data were normalized by software and fused into a data matrix,combined with partial least squares discrimination method to establish a fingerprint based on near-infrared spectrum and volatile component,Adulteration discrimination model of information fusion.The difference between the predicted value of the model and the real value artificially assigned is less than the set prediction threshold.The model determination coefficient and root mean square error show that the model has good sensitivity and stability.Comparing the models established based on single source data of near-infrared fingerprint analysis and volatile component fingerprint analysis,the amount of data analysis of the model after data fusion is small and has certain specificity.In theory,it can realize the research on the identification of Wuchang Daohuaxiang rice with binary adulteration. |
PDF Full Text Request