Study On Microwave Fixation Technology Of Tea Based On Machine Vision

Fixation is the most critical process in tea processing and plays a decisive role in the quality of tea.Microwave fixation technology has been widely used because of its advantages of rapid heating up,energy saving and high efficiency.However,the traditional fixation effect mainly relies on the personal sense evaluation of tea makers,which was highly subjective and labor-intensive.In this paper,machine vision was used to replace artificial sense,and a tea microwave fixation test system based on machine vision was established.The influence of the fixation temperature on the physical features,internal quality and sensory quality of tea leaves were studied.By establishing mathematical models,the application of machine vision technology in the rapid detection of moisture content of tea leaves was discussed.By establishing the relationship between physical features and internal quality,a fuzzy control strategy of fixation temperature based on machine vision was designed to realize the adaptive control of fixation temperature during the tea microwave fixation process.The main research contents are as follows:（1）A tea microwave fixation test system based on machine vision was established.The microwave fixation unit can be controlled by the upper computer,and its power can be adjusted continuously;the temperature detection unit detects the temperature signal of tea leaves in real time by the optical fiber sensor;the weight detection unit detects the weight of the tea leaves in real time by an electronic balance;the visual detection unit collects the tea images in real time by an industrial camera;the PC control unit completes the conversion and feedback control of temperature,the calculation of moisture content,image processing and feature extraction and data storage online.The temperature control error of the system was within±1℃,and the temperature error of different positions in the microwave cavity was within±1.18℃.（2）The influence of temperature（60℃,70℃,80℃）on the RGB,Lab color features and the contrast,dissimilarity,entropy,homogeneity,correlation and energy texture features of tea leaves during the fixation process,on the moisture content and water loss rate of tea leaves during the fixation process,on the polyphenol content,amino acid content,chlorophyll content and PPO activity of tea leaves after fixation and on the sensory quality of the finished tea leaves were studied.The research showed that the color and texture feature values changed regularly,and the higher the fixation temperature,the better the color formation and the expression of texture.The water loss rate could be divided into four categories:rapid increase,constant speed,rapid decline and slow decline,and the higher the fixation temperature,the faster the water content decreases,the greater the water loss rate and the shorter the time consuming.The higher the fixation temperature,the higher the polyphenol content and the phenol-to-ammonia ratio,the lower the amino acid content,chlorophyll content and PPO activity,the better the appearance,aroma and leaf bottom of the finished tea leaves.（3）The correlation analysis method,forward selection method,forward selection method combined with correlation analysis method were used to screen physical features,and the BP,ELM and GA-BP mathematical models were established respectively with the screened feature variables as input,as well as the PLSR model with all features as input was established to discusse the application of machine vision technology in the rapid detection of the moisture content of tea leaves.The research showed that the performance of the optimal PLSR-6 model in PLSR models was relatively low.The combination method could effectively improve the detection performance of models compared with the single screening method.Under the same screening method,the performance of the BP,ELM and GA-BP models improved successively.Taking the G value,B value,correlation and|a|value screened out by the combination method as input,GA-BP model had the best performance,with R_C² and R_P² were 0.9947 and 0.9942,respectively,RMSEC and RMSEP were 0.5503 and 0.5772,respectively.（4）A fuzzy control strategy of the fixation temperature based on machine vision was designed,and the adaptive control of fixation temperature during microwave fixation process was realized.The resea_rch showed that the R value could be used as a judgment condition to divide the microwave fixation process into two stages:early stage and mid-late stage.When?R≥0,it was in the early stage,and when?R<0,it was in the mid-late stage.In the early stage of fixation,the|a|value and|?a|value were used to control temperature to passivate the PPO activity and reduce the damage to chlorophyll.And in the mid-late stage,the R value and the dissimilarity change value were used to control temperature to ensure the formation of color,the retention of flavor substances and the transformation of aroma substances.Under the fuzzy control scheme,the content of tea polyphenols of tea leaves after fixation was 16.23%,the content of amino acids was 6.67%,the ratio of tea polyphenols to amino acids was 2.43,the content of chlorophyll was 9.33 mg/g,the PPO activity was 4.98 U/m L,the sensory score of the finished tea leaves was 91.5,and both intrinsic and sensory quality were improved.