基于光谱和气味特征信息的霉变大米无损检测

摘要

为快速鉴定大米霉变程度,本研究运用近/中红外光谱和电子鼻分析技术,建立了大米有害霉菌侵染种类与霉变程度的同步识别方法.首先,将4种谷物中常见有害霉菌分别接种在灭菌大米样品上,将样品于28℃和80％ RH环境条件下储藏10 d.其次,样品从接种霉菌起,选取时间节点0、2、4、7、10 d,获得其近/中红外光谱和电子鼻气味特征信息.结果显示,受不同霉菌侵染大米样品的光谱和气味整体信息存在差异,到储藏后期差异更加显著.结合主成分分析和线性判别分析法,近/中红外光谱和电子鼻对受不同霉菌侵染大米样品的整体识别率分别为86.0％、86.0％、92.0％.大米霉变程度随储藏时间逐渐加深,近/中红外光谱和电子鼻对感染单一霉菌样品霉变程度的判别正确率达97.5％、98.75％、100％,多种霉菌感染的判别正确率为80.0％、87.5％、95.0％.结果表明,利用光谱和气味特征信息实现大米霉变的快速检测具有可行性,电子鼻在霉变大米特征挥发性气味的识别方面更具优势.%To rapidly identify moldy degree of rice,classification models were adopted for the detection of harmful fungi species and the degree of infection based on near/mid infrared spectroscopy and electronic nose techniques.Firstly,the clean and sterile rice samples were inoculated with 4 different spore suspensions of Aspergillus spp.Then,all infected and control samples were stored at 28 ℃ and 80％ relative humidity (RH) for 10 d until all peanut samples were covered with a mass of fungi.Finally,near/mid spectral and volatile odor information from rice samples stored for 0,2,4,7 and 10 d were collected.The results showed that:the differences in spectral and odor information of rice samples infected by different fungi species did exist.Combined with principal component analysis (PCA) and linear discriminant analysis (LD)A),the correct rate obtained by NIR,MIR and E-nose for the classification of samples contaminated by different fungal species were 86.0％,86.0％ and 92.0％,respectively.The degree of mildew was intensified during the storage process.The average correct classification accuracy of the storage time (mildew degree) was found to be 97.5％,98.75％,100％ for samples infected by one fungi spice,and 80.0％,87.5％,95.0％ for samples infected by the four fungi spices for NIR,MIR and E-nose,respectively.The results indicated that moldy rice could be detected by the usage of spectral and volatile odor information.E-nose had more advantages in identifying the volatile odor of moldy rice.