• EI
  • Scopus
  • 食品科学与工程领域高质量科技期刊分级目录第一方阵T1
  • DOAJ
  • EBSCO
  • 北大核心期刊
  • 中国核心学术期刊RCCSE
  • JST China
  • FSTA
  • 中国精品科技期刊
  • 中国农业核心期刊
  • CA
  • WJCI
  • 中国科技核心期刊CSTPCD
  • 中国生物医学SinoMed
中国精品科技期刊2020
林云,欧阳璐斯,赖燕华,等. 基于霉菌酵母测试片和NIR技术快速鉴别霉变烟草最优势霉菌种类[J]. 食品工业科技,2021,42(23):280−286. doi: 10.13386/j.issn1002-0306.2021030194.
引用本文: 林云,欧阳璐斯,赖燕华,等. 基于霉菌酵母测试片和NIR技术快速鉴别霉变烟草最优势霉菌种类[J]. 食品工业科技,2021,42(23):280−286. doi: 10.13386/j.issn1002-0306.2021030194.
LIN Yun, OUYANG Lusi, LAI Yanhua, et al. Rapid Identification of the Most Dominant Mold on Moldy Tobacco Leaves Based on Rapid Test Strips and NIR Technology[J]. Science and Technology of Food Industry, 2021, 42(23): 280−286. (in Chinese with English abstract). doi: 10.13386/j.issn1002-0306.2021030194.
Citation: LIN Yun, OUYANG Lusi, LAI Yanhua, et al. Rapid Identification of the Most Dominant Mold on Moldy Tobacco Leaves Based on Rapid Test Strips and NIR Technology[J]. Science and Technology of Food Industry, 2021, 42(23): 280−286. (in Chinese with English abstract). doi: 10.13386/j.issn1002-0306.2021030194.

基于霉菌酵母测试片和NIR技术快速鉴别霉变烟草最优势霉菌种类

Rapid Identification of the Most Dominant Mold on Moldy Tobacco Leaves Based on Rapid Test Strips and NIR Technology

  • 摘要: 为对烟草的最优势霉菌种类进行更具时效性的监控,本文首次提出一种霉菌酵母测试片结合NIR技术快速鉴别霉变烟叶上最优势霉菌种类的方法。将霉变烟叶上的霉菌用生理盐水稀释成不同梯度,并将稀释液制备成含菌测试片。应用NIR技术对培养后的含菌测试片进行光谱采集,获得具有不同种类最优势霉菌特征信息的近红外光谱数据。利用离散小波变换对光谱进行预处理,探究基于不同的小波基函数和小波分解层数的光谱预处理效果,运用随机森林算法构建不同最优势霉菌种类的鉴别模型。训练集识别正确率达98.25%,测试集预测正确率达99.30%,模型分类性能良好,能够用于识别常见的烟草最优势霉菌种类。本方法操作步骤简单,检测速度快,预测准确率高,为特定环境微生物种属鉴定提供了新的思路。

     

    Abstract: In order to identify the species of the most dominant mold on moldy tobacco leaves rapidly, a method of rapid identification of the most dominant mold on moldy tobacco leaves based on rapid test strips and near infrared spectroscopy (NIR) technology was proposed for the first time in this paper. The mold on moldy tobacco leaves was diluted into different gradients by physiological saline to prepare rapid test strip samples with mold. NIR technology was applied for collecting the spectra of rapid test strips with the messages of different species of the most dominant mold. The spectra were pretreated by the discrete wavelet transformation (DWT) method, which was optimized by using different wavelet basis function and decomposition layers. The training set of the most dominant mold was modeled based on wavelet coefficients by using random forest (RF). The correct recognition rate of training dataset was 98.25%, while that for testing dataset was 99.30%. The model had good classification performance and could be used to identify the most dominant mold species in tobacco. Application of rapid test strips combined with NIR technology could identify the species of the most dominant mold on moldy tobacco leaves rapidly and accurately, providing a new train of thought for microorganism identification.

     

/

返回文章
返回