Study On The Judgment Method Of The Degree Of Sealed Yellowing Of Yellow Tea

Yellow tea is a large type of tea unique to China,and its initial procedures,includes fixing,sealed yellowing and drying,in which sealed yellowing is the key process for yellow tea.The correct grasp of the sealed yellowing degree determines the quality of yellow tea directly.In production,the control of the sealed yellowing degree of yellow tea mainly depends on the experience of the tea maker,lacking scientificity and stability.In this experiment,the process of sealed yellowing were used as the research object,and the change rules of the sensory quality and main biochemical components of yellow tea with the extension of sealed yellowing time and the change rules of the electrochemical parameters(p H,ORP and conductivity)and color of sealed yellowing leaf with the extension of sealed yellowing time were systematically investigated with sensory evaluation,biochemical test detection,electrochemical analysis and computer vision technology,and some models for judging the degree of sealed yellowing were established based on the electrochemical parameters and color characteristics.The main research results are as follows:1.Research on the change rules of sensory quality and main biochemical composition of yellow tea with the extension of sealed yellowing timeWith the prolongation of sealed yellowing time,the sensory quality of yellow tea showed a trend of rising first and then falling.The sensory quality of sealed yellowing 6 h tea sample was the best was the best.The contents of tea polyphenols,catechins,chlorophyll a,chlorophyll b and total chlorophyll all showed a downward trend.The phenol-ammonia ratio decreased first and then increased.The content of caffeine didn't change significantly with the extension of sealed yellowing time.2.Research on the change rules of electrochemical parame ters and color of sealed yellowing leaf with the extension of sealed yellowing timeBoth of the changes of hot bubble p H and cold bubble p H showed a downward trend.Both of the changes of hot bubble ORP and cold bubble ORP showed an upward trend,while both the hot bubble conductivity and cold bubble conductivity didn't change significantly.The r,which represents the proportion of red,showed an upward trend generally.The g,which represents the proportion of green,showed a downward trend generally.Hue showed downward trend generally.Saturation and Value had no significant changes trend.3.Research on judging the degree of sealed yellowing based on electrochemical parametersBased on the electrochemical parameters,we established Fisher's discriminant model and SVM model to judge the degree of sealed yellowing and the regression model to predict the contents of tea polyphenols,catechins and theaflavin.The results show both of the Fisher discriminant model and SVM model had good effects.The average accuracies of training set and prediction set of Fisher discriminant model were 93.33%and 86.66%,respectively.The average accuracies of training set and prediction set of SVM model were92.22%and 86.66%,respectively.Both of the regression models of tea polyphenols and catechins had good prediction effects.Their R~2 were 0.814 and 0.956 and their average relative errors were 3.84%and 4.13%,respectively.The regression models of theaflavin had a bad prediction effect.Its R~2 was 0.748 and its average relative error was 39.33%.4.Research on judging the degree of sealed yellowing based on color characteristicsBased on the color characteristics,we established Fisher's discriminant model and SVM model to judge the degree of sealed yellowing,and the regression model to predict the contents of chlorophyll a,chlorophyll b,total chlorophyll and theaflavin.The results show the effect of Fisher discriminant model was average and the effect of SVM model was good.The average accuracies of training set and prediction set of Fisher discriminant model were 86.66%and 76.66%,respectively.The average accuracies of training set and prediction set of SVM model were 92.22%and 90%,respectively.All of the regression models of chlorophyll a,chlorophyll b and total chlorophyll had good prediction effects.Their R~2 were 0.948,0.911 and 0.949 and their average relative errors were 1.05%,4.66%and 1.93%,respectively.The regression models of theaflavin had a bad prediction effect.Its R~2 was 0.783 and its average relative error was 12.34%.