摘要
采用由6个金属氧化物气敏传感器组成阵列的电子鼻对2个等级的信阳毛尖茶进行检测,并通过主成分分析(PCA)、判别分析(LDA)和BP神经网络对数据进行分析和识别。PCA和LDA结果显示,可以将2个等级的茶叶完全区分开。采用3层BP神经网络对数据矩阵进行茶叶等级的定量预测,预测结果平均相对误差为1.16,最大相对误差为13.32。研究结果表明,供试气敏传感器阵列对信阳毛尖茶等级的检测具有很高的定量分析精度。
Xingyang Maojian tea grade were measured by the gas sensor s array which was composed of six metal oxide semiconductor gas sensors. Principal component analysis （PCA）,lineardiscriminant analysis （LDA） and BP network were used in the data analysis and pattern recognition. The results obtained prove that the electronic nose can discriminate successfully different grade of tea using PCA and LDA analysis. A feed forward artificial neural （BP）network with three layers to predict the tea grade achieving average relative error of 1.16 with max relative error of 13.32. The results show that gas sensor array could predict the tea grade with a high accuracy.
出处
《河南农业科学》
CSCD
北大核心
2010年第4期36-38,共3页
Journal of Henan Agricultural Sciences
基金
河南省教育厅自然科学研究计划项目(2009B210017)
关键词
传感器阵列
信阳毛尖茶
等级
Gas sensor array
Xinyang Maojian tea
Grade
