Study On Olfactory And Gustatory Dual Channel Perception Method For Green Tea Quality And Flavor

The aroma and taste of tea play an important role in its quality and flavor.At present,the flavor detection methods can not meet the needs of rapid detection of green tea quality and flavor.Therefore,in this study,electronic nose and taste visualization technology were used to obtain the olfactory and taste information of green tea to establish olfactory and gustatory dual channel perception methods for green tea quality and flavor;and quantitative evaluation models were established with artificial sensory evaluation as reference;furthermore,the olfactory and taste sensory interactions were introduced to explore the effect of taste perception affected by olfaction on the quality and flavor of green tea.The main contents of this study are as follows:1.Physicochemical detection and artificial sensory evaluation of green tea samplesAccording to the national standard,the main taste substances in green tea were detected to provide references for establishing the quantitative prediction models of taste substances in green tea in the later stage;in accordance with the evaluation method in the national standard,the quality and flavor of green tea were evaluated,and the sensory score of the sample was obtained,which was used to establish the prediction model of sensory score of green tea quality and flavor.2.Research on olfactory quality and flavor detection method of green tea based on electronic nose technologyThe characteristic signals of green tea samples were obtained by electronic nose technology,and the discrimination models of olfactory flavor quality of green tea were established.According to the collected aroma information of green tea,the data were preprocessed by standardization methods.K-nearest neighbor(KNN)method and support vector machine(SVM)method were used to establish pattern recognition models to distinguish different types of green tea.The discriminant accuracy rates of KNN model for training set and prediction set were 97.78% and 96.67% respectively,and that of SVM model were 100%.3.Research on taste quality flavor detection method of green tea based on taste visualization technologyThe colormetric sensor arroy was constructed according to the selected color sensitive materials,and then the taste visualization technology was used to collect the taste information of green tea.The previous 10 principal components were used as characteristic value to establish KNN and SVM recognition models respectively to distinguish the quality and flavor of green tea samples.The accuracy rates of the two models were 67.78% and 83.33%,respectively.The quantitative prediction models of tea polyphenols,catechins,caffeine and free amino acids were established by partial least squares(PLS)and back propagation neural network(BPNN).The correlation coefficients(Rp)of PLS model for the four taste substances were 0.7969,0.8905,0.9078 and 0.7762,and the root mean square errors of the prediction set(RMSEP)were 1.0900,0.8310,0.2980 and 0.1940,respectively.The Rp of the BPNN prediction model for the four taste compounds was 0.8809,0.9568,0.9336 and 0.8640,and the RMSEP was 4.5115,2.5249,1.4136 and 0.8973,respectively.The effect of BPNN model was better than that of PLS model.The results showed that the taste visualization technology could achieve the quantitative prediction of the content of characteristic taste substances in green tea.4.Research on sensory score detection method of green tea based on sensor information fusion technologyThe dual channel perception method of olfactory and gustation of green tea quality and flavor was studied by information fusion technology.The electronic nose and taste visualization were used to obtain the odor and taste characteristics of green tea,and the quantitative prediction models of green tea sensory score were established.The quantitative prediction models of green tea aroma score were established by using the characteristic signal of electronic nose.The Rp and RMSEP of BPNN model were0.8533 and 3.4984 respectively.According to the characteristic signals of taste visualization sensor,the quantitative prediction models of green tea taste score were established.The Rp and RMSEP of BPNN model were 0.8322 and 5.4308,respectively.Based on the fusion information of electronic nose and taste visualization,the quantitative prediction models of green tea overall score were established.The Rp and RMSEP of BPNN model were 0.8980 and 13.4528,respectively.The results showed that the prediction accuracy of sensor information fusion technology was better than that of single technology.5.Research on discrimination of tea quality flavor by colorimetric sensor array based on olfactory-taste interactionIn the process of food sensory evaluation,the interaction between smell and taste has an important impact on the evaluation of green tea flavor.Therefore,this study introduced the interaction between olfactory and taste in the evaluation of different quality and flavor of green tea.The characteristic signals of colorimetric sensor array of taste affected by olfaction were extracted as the input of the models,and the discrimination models of different quality and flavor green tea were established by KNN and SVM methods.The discriminant accuracy rates of training set and prediction set of KNN model were 90% and 82.22%,respectively,and that of SVM model were94.44% and 91.11%.The results showed that it was feasible to evaluate the quality and flavor of green tea by olfactory and gustatory dual channel perception methods based on electronic nose and taste visualization technology.The discrimination method of colorimetric sensor array based on olfactory and gustatory interaction could accurately identify the quality flavor of green tea.The results of this study could be used in the evaluation of green tea quality and flavor,and also provide references for other food quality flavor evaluation.