基于电子鼻技术对不同类型洋葱提取液的识别

Identification of different types of onion extracts by electronic nose technology

中文关键词:  电子鼻  洋葱  植物学性状  费舍尔判别  反向传播神经网络

英文关键词:electronic nose  onion extracts  botany traits  Fisher discrimination  back propagation neural network

中文摘要:

      目的 研究电子鼻技术对不同类型洋葱提取液的快速识别。方法 以云南、甘肃、安徽、四川、山东、江苏的紫皮洋葱, 甘肃、吉林、云南的黄皮洋葱和新疆的白皮洋葱为实验对象, 测定植物学性状后提取洋葱的提取液, 运用电子鼻检测分析提取液的挥发性物质, 采用费舍尔(Fisher)判别和反向传播神经网络(back propagation neural network, BPNN)建立预测模型。结果 电子鼻的10个传感器对不同类型的洋葱提取液的响应值有显著性差异(P<0.05), Fisher判别模型和BPNN模型均可有效地识别不同类型的洋葱提取液, BPNN对训练集和检验集的识别正确率分别为100%和98.3%, Fisher判别对训练集和检验集的识别正确率分别为96.1%和92.8%。电子鼻技术结合BPNN更适合不同类型洋葱提取液的识别。结论 电子鼻技术结合BPNN可以识别不同类型的洋葱提取液, 为果蔬保鲜的应用开发提供了理论依据和技术支持。

英文摘要:

      Objective To study the rapid identification of different types of onion extracts by electronic nose technology. Methods Purple onions grown in Yunnan, Gansu, Anhui, Sichuan, Shandong and Jiangsu regions plus yellow onions grown in Gansu, Jilin and Yunnan regions plus white onion from Xinjiang region, were used for materials, after the botany traits were investigated, onions were extracted and tested by electronic nose, identification models were established by using Fisher discrimination and back propagation neural network (BPNN). Results The responses of 10 sensors to different types of onion extracts were significantly different (P<0.05). Fisher discriminant model and BPNN model could effectively identify different types of onion extracts, the recognition accuracies of BPNN for training set and test set were 100% and 98.3% respectively, and the recognition accuracies of Fisher discriminant for training set and test set were 96.1% and 92.8% respectively. Electronic nose technology combined with BPNN was more suitable for the identification of different types of onion extracts. Conclusion Electronic nose technology combined with BPNN can identify different types of onion extracts, which can provide theoretical basis and technical support for application development of fruit and vegetable preservation.