电子鼻传感器阵列优化与谷物霉变程度的检测  

Optimized of Sensor Array and Detection of Moldy Degree for Grain by Electronic Nose

作　　者：张红梅[1] 王俊[1] 叶盛[1] 于慧春[1] 田晓静[1] 

机构地区：[1]浙江大学生物系统工程与食品科学学院,杭州310029

出　　处：《传感技术学报》2007年第6期1207-1210,共4页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金资助(30571076);国家教育部新世纪人才支持计划资助(NCET-04-0544)

摘　　要：研制一套适合对谷物霉变进行检测的电子鼻系统,对6个霉变程度的稻谷进行了检测.主成份(PCA)分析结果显示6个霉变程度的稻谷是可以区分的,对前三个主成分的载荷因子进行分析去掉冗余传感器,PCA分析结果显示去掉冗余传感器并不影响分类结果.最后用BP神经网络对所得的数据进行模式识别,30组测试样本对霉变菌落总数的预测结果平均相对误差为1.01030,最大相对误差为3.94257.实验结果表明系统对稻谷霉变程度的检测具有很高的分析精度.An electronic nose system was developed for the evaluation of moldy grain. The system has detected six different moldy degrees of rice. An initial eight sensors was simplified to six after a PCA（Principal component analysis） analysis detected redundancy between of the sensors, Data processing and classification performed by PCA analysis showed no significant performance differences between the complete and reduced sensor array. A feedforward artificial neural network with test data predict the aerobic bacterial count achieving average relative error of 1. 010 30, max relative error of 3. 942 57. Our results show that Electronic Nose system could predict the moldy rice with a high accuracy.

关 键 词：电子鼻 传感器阵列 主成分分析 BP网络 菌落总数 

分 类 号：TP212.6[自动化与计算机技术—检测技术与自动化装置]