摘要
茶叶加工过程的状态参数监测是自动化生产的重点和难点。针对茶叶杀青过程中的状态监测,利用自制的电子鼻,在优选传感器阵列的基础上,对杀青过程香气开展连续抽样监测,基于费舍尔判别法研究茶叶加工过程香气数据与茶叶加工状态的关系,构建茶叶加工状态的判别模型,试验表明建模组的回判准确率将近99%,验证组的判别准确率也达到了91.65%,对茶叶杀青过程状态的在线监测与控制具有重要意义。
The state monitoring of tea processing is the key and difficult point of automatic production.In view of the condition monitoring in the status of tea de-enzyming,based on the self-made electronic nose and the optimized sensor array.This paper carried out the continuous sampling monitoring of the aroma in the process of tea de-enzyming,studied the relationship between the aroma data and the tea de-enzyming state based on the fisher discriminant method,and construced the discriminant model of tea de-enzyming state.The experiment showed that the discriminating accuracy of the training group was nearly 99%,and the accuracy rate of the validation group reached nearly 91.65%,which was important for the online monitoring and control of the state of tea green process.
作者
何维
王进
黄九明
陆国栋
唐小林
李文萃
HE Wei;WANG Jin;HUANG Jiuming;LU Guodong;TANG Xiaolin;LI Wencui(State Key Laboratory of Fluid Power&Mechatronic Systems,Zhejiang University,Hangzhou,Zhejiang 310027,China;Hangzhou Tea Research Institute,China COOP,Hangzhou,Zhejiang 310016,China)
出处
《农产品加工》
2020年第24期53-58,共6页
Farm Products Processing
基金
浙江省重点研发计划项目(2017C02007)。
关键词
绿茶杀青
加工状态
费舍尔判别法
在线监测
tea de-enzyming
processing state
fisher discriminant
online monitoring
