基于电子鼻技术的山核桃陈化指标预测模型研究

Detection Models of Aging Index of Walnut(Carya Cathayensis Sarg) Based on Electronic Nose Technology

通过分析电子鼻采集的多传感器信号与自然陈化指标间的关联特征,建立基于电子鼻检测的山核桃陈化指标预测模型,实现对其陈化指标的预测,快速区分辨别山核桃贮藏年限。采用电子鼻系统对4个贮藏年限的山核桃进行了检测,PCA分析可以将不同贮藏年限的山核桃完全区分开,且效果较好。进而利用主成分回归方法建立了基于气敏传感器阵列数据的山核桃陈化指标预测模型,并用预测集对模型进行验证。结果表明碘价、过氧化值、酸价和茴香胺含量预测值与实测值之间的相关系数较高,预测标准误差和平均误差百分比较小,它们分别为0.88%、6.79%和6.8%;0.78%、1.23%和2.3%;0.82%、0.127%和1.58%;0.81%、0.61%和0.76%,其预测值能够精确地反映不同贮藏年限山核桃的陈化指标,因此,电子鼻技术可以用于不同贮藏年限的山核桃陈化指标含量检测。

A prediction model was established to rapidly distinguish storage years of walnuts(Carya Cathayensis Sarg)in this study.This model was based on the correlation between multi-sensor signals collected by electronic nose and natural aging index.According to the PCA analysis of data detected by electronic nose,we found that the walnuts with four storage years were completely separated.And then,walnut aging indexes prediction models were established by using principal component regression based on the gas sensor array data.The predicted and the measured value including iodine value,oxidation value,acid value and anisidine content,were with high correlation coefficients(0.88,0.78,0.82,and 0.81,respectively),small prediction standard errors(6.79,1.23,0.127,and 0.61,respectively)and percentage of average error(6.8%,2.3%,1.58%,and 0.76%,respectively).Therefore,electronic nose technology can be used to identify the storage years of walnut.