摘要
为探索绿豆产地鉴别可行性,实现对来自白城、杜尔伯特、泰来、泗水绿豆原产地的快速鉴别。研究运用近红外光谱技术,并结合PLS-DA法(偏最小二乘判别分析)进行建模,对4个产地的120份绿豆粉末样品进行近红外光谱的扫描,分别采用不同光谱预处理方法[标准正态变换(SNV)、多元散射校正(MSC)、矢量归一化与导数处理等]最终确定矢量归一化+MSC建立的模型最稳定,建模波长为12 000~4 000 cm-1全波长。主成分分析提取3个有效主成分(主成分1贡献率为52.44%,主成分2贡献率为30.16%,主成分3贡献率为9.57%),其累计贡献率达到92.17%。用预测样本集进行模型的验证,白城、杜尔伯特、泰来、泗水4个产地的预测正确率分别为100%、80%、80%和100%。预测结果达到80%以上,初步认定近红外光谱分析技术可用于绿豆产地溯源研究。
In order to explore the feasibility of mung bean origin identification,rapid identification of the origin of mung bean from Baicheng,Dumeng,Tailai and Sishui was realized.Near-infrared spectroscopy with PLS-DA method(partial least squares discriminant analysis)was used for modeling,and 120 samples of mung bean powder from four producing areas were scanned by near-infrared spectrum.Different spectral preprocessing methods〔standard normal transformation(SNV),multiple scattering correction(MSC),vector normalization and derivative processing,etc.〕were used to determine that the model established by vector normalization+MSC was the most stable,and the modeling wavelength was 12000~4000 cm-1 full wavelength.Three effective principal components were extracted by principal component analysis(contribution rate of principal component 1 was 52.44%,that of principal component 2 was 30.16%,and that of principal component 3 was 9.57%),and their cumulative contribution rate reached 92.17%.The prediction accuracy of Baicheng,Dumeng,Tailai and Sishui is 100%,80%,80%and 100%,respectively.The predicted results reached more than 80%,and it was preliminarily recognized that near infrared spectroscopy could be used to trace the origin of mung bean.
作者
陈明明
邱彦超
宋妍
杨斯琪
左锋
钱丽丽
Chen Mingming;Qiu Yanchao;Song Yan;Yang Siqi;Zuo Feng;Qian Lili(College of Food Science,Heilongjiang Bayi Agricultural University,Daqing 163319;National Coarse Cereals Engineering Research Center;Key Laboratory of Agro-products Processing and Quality Safety of Heilongjiang Province)
出处
《黑龙江八一农垦大学学报》
2024年第1期49-54,共6页
journal of heilongjiang bayi agricultural university
基金
黑龙江省杂粮产业技术协同创新体系杂粮质量溯源技术岗、黑龙江省特色学科资助项目(黑教联[2018]4号)
国家重点研发计划项目(杂粮食品精细化加工关键技术合作研究及应用示范:2018YFE0206300)。
关键词
绿豆
产地鉴别
近红外光谱
主成分分析
偏最小二乘判别分析
mung bean
origin identification
near-infrared spectrum
principal component analysis
partial least squares discriminant analysis
