摘要
对近红外光谱分析技术鉴别不同品牌的方便面进行了研究。选取三个代表性品牌,每个品牌各制45个样(其中40个样品用于建模,5个样品用于预测),通过比较不同光谱预处理方法及光谱范围,得到最优模型。结果表明,在8350～4190cm-1光谱范围内,采用矢量归一法对光谱数据进行预处理,再分别进行聚类分析与定性分析,得到的模型的品牌识别率均为100%,比较两种方法建立模型的效率,聚类分析比定性分析更加快速。说明应用近红外光谱分析技术能够快速、有效的鉴别方便面品牌,为方便面品牌的鉴别提供了一种新方法。
Near-infrared Spectroscopy technology was used to distinguish three brands of instant noodles.In the experiment,45 samples were used for each brand of instant noodles,among which 40 were used for modeling and 5 for brand prediction.By comparing several preprocessing procedures and wavelength ranges,the optimal models could be obtained.The result showed that in the range of 8350-4190 cm-1 with Vector Normalization for pretreatment,identification rate of the instant noodles of two brand was 100% using cluster analysis and qualitative analysis.The efficiency of the models made by cluster analysis was higher than that by qualitative analysis.This method was rapid and practicable,suitable for the discrimination of brands of instant noodles.
出处
《现代食品科技》
EI
CAS
2010年第10期1153-1156,共4页
Modern Food Science and Technology
基金
陕西省科技攻关项目(2007ZDKG-01)
西北农林科技大学人才基金项目
关键词
近红外光谱
方便面
品牌
聚类分析
near-infrared spectroscopy
instant noodles
brand
cluster analysis
