欢迎访问《中国粮油学报》编辑部网站！

赵天霞,沈飞,周曰春,刘潇,方勇,李彭,裴斐,邢常瑞.小麦霉菌侵染程度电子鼻快速检测方法的初步研究[J].中国粮油学报,2019,34(6):134-140

小麦霉菌侵染程度电子鼻快速检测方法的初步研究

Preliminary Study on Rapid Detection of Fungal Infection in Wheat Based on Electronic Nose

投稿时间：2018-09-04 修订日期：2018-11-29

DOI：

英文关键词:wheat fungal infection mildew degree electronic nose rapid detection

基金项目:国家自然科学基金项目（面上项目，重点项目，重大项目）

作者	单位	邮编
赵天霞	南京财经大学	210023
沈飞^*	南京财经大学	210023
周曰春	南京灵山粮食储备库有限公司
刘潇	南京财经大学
方勇	南京财经大学
李彭	南京财经大学
裴斐	南京财经大学
邢常瑞	南京财经大学

摘要点击次数: 541

全文下载次数: 53

中文摘要:

本研究利用电子鼻气体传感技术，初步建立了小麦霉菌侵染程度定性定量同步分析方法。小麦样品经辐照杀菌后接种5种谷物中常见有害霉菌，于85%相对湿度和28℃的环境中储藏至重度霉变。在样品储藏的不同阶段，选取时间节点0、1、3、5和7 d采集其电子鼻气味响应信息，建立了其响应信号和霉菌侵染程度的相关关系模型。结果显示，依据带菌量的不同，基于电子鼻信号的主成分分析法（PCA）可成功区分未霉变（< 2.7 log CFU/g）、轻度霉变（2.7～4 log CFU/g）与重度霉变（> 4 log CFU/g）的小麦样品；线性判别分析（LDA）对受单一霉菌侵染的小麦样品霉变程度的识别率达90.0%以上，对所有小麦样品的识别率达84.0%。偏最小二乘回归模型（PLSR）对小麦菌落总数的模型决定系数（Rp2）和预测误差（RMSEP）及相对分析偏差（RPD）分别为0.852，0.504 log CFU/g和2.30。结果表明，利用电子鼻技术实现小麦霉菌侵染程度的快速识别是可行的。下一步应不断补充不同来源的小麦样品，以不断提高模型的精度和适用性。

英文摘要:

In this study, E-nose gas sensing technology was used to establish a qualitative and quantitative analysis method for harmful fungal infection detection in wheat. Wheat samples after irradiation sterilization were inoculated with five fungal species and stored at 28 °C and 85% relative humidity until seriously moldy. E-nose odor response information of samples was collected at time storage stages of 0, 1, 3, 5 and 7 d, and a correlation model was established between E-nose response signals and the degree of fungal infection. The results showed that principal component analysis (PCA) based on the E-nose signals could successfully distinguish healthy（< 2.7 log CFU/g）, moldy（2.7～4 log CFU/g）and highly moldy（> 4 log CFU/g）wheat samples according to colony counts. The correct classification accuracy of mildew degree was more than 90.0% for wheat samples contaminated with one fungi spice, and achieved 84.0% for samples infected by the five fungi spices. Colony counts in samples was predicted by PLSR and coefficient of determination for the prediction set (Rp2), root mean square error of prediction (RMSEP) and residual predictive deviation (RPD) value obtained were 0.852, 0.504 log CFU/g and 2.30, respectively. The results indicated that it is feasible to use electronic nose for rapid identification of harmful fungal infection in wheat. In next step, natural infected wheat samples and samples contaminated with other fungal strains should be incorporated to enhance the applicability of this methodology.

关闭