基于激光拉曼光谱的榛子油掺伪定量研究
Quantitative Research on Hazelnut Oil Adulteration Based on Laser Raman Spectroscopy
张凤娟 1黄敏2
作者信息
- 1. 无锡科技职业学院集成电路学院,江苏 无锡 214028
- 2. 江南大学轻工业过程先进控制教育部重点实验室,江苏 无锡 214122
- 折叠
摘要
为探究榛子油掺伪定量检测,采用便携式激光拉曼光谱仪对榛子油分别掺入玉米油、核桃油和亚麻籽油的180个掺伪样品进行拉曼光谱扫描.对采集的光谱按3∶1比例划分校准集和验证集,利用主成分分析方法进行定性分析,通过建立偏最小二乘回归校准模型实现掺伪样品的定量检测.结果显示榛子油掺伪亚麻籽油、核桃油和玉米油的三种二元掺伪样品,其相关系数分别是0.9894、0.9872、0.9688,校正均方根误差分别为是0.0037、0.0098、0.0121,预测均方根误差分别是0.0114、0.0126、0.0190,相对分析误差分别为:9.707、8.848、5.662,并对三种掺伪样品的模型参数差异性进行了合理解释.结果表明:本模型在榛子油掺伪定量检测方面具有优秀的预测性能,使用本模型可以对榛子油掺伪进行简易、快速、无损定量检测.
Abstract
This study explored the quantitative detection of hazelnut oil adulteration.A portable laser Raman spectrometer was used on 180 adulterated samples of hazelnut oil mixed with corn oil,walnut oil,and flaxseed oil.The acquired spectra were divided into calibration and validation sets in a 3∶1 ratio.Qualitative analysis was performed using principal component analysis and quantitative detection of adulterated samples was achieved by establishing a partial least-squares regression.The experiment yielded three binary pseudo samples of hazelnut oil mixed with flaxseed oil,walnut oil,and corn oil,respectively.The corresponding correlation coefficients were 0.9894,0.9872,and 0.9688;the root mean square errors on calibration were 0.0037,0.0098,and 0.0121;the root mean square errors on prediction were 0.0114,0.0126,and 0.0190;and the relative analysis errors were 9.707,8.848,and 5.662.The difference in the model parameters of the three kinds of adulterated samples was reasonably explained.The results show that the proposed system has excellent predictive performance for quantitative detection of hazelnut oil adulteration.The system can be used for simple,rapid,and nondestructive quantitative detection of hazelnut oil adulteration.
关键词
拉曼光谱/二元掺伪样品/偏最小二乘回归/定量检测Key words
Raman spectrum/binary adulterated samples/partial least squares regression/quantitative detection引用本文复制引用
出版年
2024
NETL
NSTL
万方数据