您好,欢迎访问湖南省农业科学院 机构知识库!

Portable LWNIR and SWNIR spectroscopy with pattern recognition technology for accurate and nondestructive detection of hidden mold infection in citrus

文献类型: 外文期刊

作者: Li, Pao 1 ; Su, Guanglin 1 ; Du, Guorong 4 ; Jiang, Liwen 1 ; Dong, Yiqing 1 ; Shan, Yang 1 ;

作者机构: 1.Hunan Agr Univ, Coll Food Sci & Technol, Changsha 410128, Peoples R China

2.Hunan Acad Agr Sci, Hunan Agr Prod Proc Inst, Changsha 410125, Peoples R China

3.Shaoguan Univ, Guangdong Prov Key Lab Utilizat & Conservat Food &, Shaoguan 512005, Peoples R China

4.Shanghai Tobacco Grp Co Ltd, Beijing Work Stn, Technol Ctr, Beijing 101121, Peoples R China

关键词: Nondestructive detection; Near infrared diffuse reflectance spectroscopy; Citrus; Hidden mold infection; Pattern recognition

期刊名称:MICROCHEMICAL JOURNAL ( 影响因子:4.8; 五年影响因子:4.5 )

ISSN: 0026-265X

年卷期: 2023 年 193 卷

页码:

收录情况: SCI

摘要: An accurate and nondestructive detection method of hidden mold infection in citrus was established based on portable near infrared diffuse reflectance spectroscopy (NIRDRS) and chemometric methods. Penetrability of NIRDRS light on the peel of Chunjian hybrid citrus was studied. The results show that NIRDRS light can penetrate the peel to a certain extent, while the penetrability of short-wave near infrared (SWNIR) was better than that of long-wave near infrared (LWNIR). The identification models of hidden mold infection were established with five pattern recognition methods combined with different wavelength bands. The results show that the identification models of LWNIR were much better than those of SWNIR. 100% identification accuracies of LWNIR were ob-tained with soft independent pattern classification (SIMCA), support vector machine (SVM), partial least squares discriminant analysis (PLS-DA) and the optimized pretreatment methods. In addition, the developed models were further validated by the external validation set collected one month later.

  • 相关文献
作者其他论文 更多>>