基于一维卷积神经网络结合便携式拉曼光谱特级 初榨橄榄油掺假定量分析

张焕俊¹, 戴臻² , 费洪晓

PDF(1930 KB)
光散射学报 ›› 2024, Vol. 36 ›› Issue (4) : 436-444. DOI: 10.13883/j.issn1004-5929.202404009

基于一维卷积神经网络结合便携式拉曼光谱特级 初榨橄榄油掺假定量分析

作者信息 +

Quantitative analysis of adulteration in extra virgin olive oil based on one-dimensional Convolutional neural network and portable raman spectroscopy

Author information +
History +

摘要

针对造假手段的不断提升的现状及廉价的橄榄果渣油很可能成为特级初榨橄榄油掺假的潜在原材 料等问题。因此,本研究重点围绕深度学习算法辅助非接触式无损伤光谱检测技术量化特级初榨橄榄油的掺 假行为。将过期橄榄果渣油和特级初榨橄榄油按不同体积比例混合,从而制备出不同浓度的掺假混合油品。 使用785 nm 便携式拉曼光谱仪对这些混合油品进行拉曼光谱采集,并结合一维卷积神经网络算法建立掺假 量化模型。采用密度泛函理论基于B3LYP/6-31+G(d,p) 基组计算亚油酸分子的理论振动光谱,以进一步 解析特级初榨橄榄油的拉曼光谱。结果表明,基于具有深度结构的前馈神经网络与785nm 便携式拉曼光谱 技术联用的技术方案是定量分析植物油掺假的有力工具,80个混合油品的4000条光谱数据量化模型的决定 系数均优于0.97,其中评价模型测试集定量分析的决定系数达到了0.9704,均方根误差小于0.0499。该技 术在快速评估特级初榨橄榄油掺假方面具有很好的应用潜力,为规范国内橄榄油市场和维护消费者合法权益 提供了一种有益的参考方案。

Abstract

Faced with the continuous improvement of counterfeiting methods,cheap olive pomace oil will probably become a potential raw material for adulterating extra virgin olive oil.Therefore,this study focuses on the deep learning algorithm-assisted non-contact,non- destructive spectral detection technology to quantify the adulteration behavior of extra virgin olive oil.Mix expired olive pomace oil and extra virgin olive oil in different volume propor- tions to prepare different adulterated,mixed oil concentrations.The 785 nm portable Raman spectrometer was used to collect the Raman spectra of these mixed oils,and the quantitative analysis model of adulteration was established by combining the one-dimensional convolu- tional neural network algorithm.The density functional theory B3LYP/6-31+G(d,p) basis set  was  used  to  calculate  the  Raman  spectrum  of  linoleic  acid  molecules  to  further  analyze  the Raman   spectrum   of   extra   virgin   olive   oil.The   experimental   results   show   that   the   technical solution  based  on  combining  deep  structured  feedforward  neural  networks  and  785  nm  porta- ble  Raman  spectroscopy  technology  is   a  powerful  tool   for  quantitative   analysis  of  plant  oil   a- dulteration.The  decision  coefficients  of  4000   spectral  data  quantitative  models  from  80  mixed oil  products   are   all   better   than   0.97,and   the   decision   coefficient   of  quantitative   analysis   in the evaluation model test set reaches 0.9704,with a root mean square error less than

0.0499.This   technology   has   great   application   potential   in   quickly   evaluating   the   adulteration of  extra  virgin  olive  oil,providing  a  beneficial  reference   scheme   for  regulating  the  domestic olive   oil   market    and   safeguarding    consumers'legitimate   rights    and   interests.

关键词

一维卷积神经网络 / 便携式拉曼光谱 / 密度泛函理论 / 特级初榨橄榄油 / 橄榄果榨油 / 掺假量化

Key words

One-dimensional / Convolutional / neural / network / Portable / Raman / spectroscopy / Density functional theory / Extra virgin olive oil / Olive pomace oil / Adulteration quantization

引用本文

导出引用
张焕俊¹, 戴臻² , 费洪晓. 基于一维卷积神经网络结合便携式拉曼光谱特级 初榨橄榄油掺假定量分析. 光散射学报. 2024, 36(4): 436-444 https://doi.org/10.13883/j.issn1004-5929.202404009
ZHANG Huanjun¹, DAI Zhen², FEI Hongxiao³. Quantitative analysis of adulteration in extra virgin olive oil based on one-dimensional Convolutional neural network and portable raman spectroscopy. Chinese Journal of Light Scattering. 2024, 36(4): 436-444 https://doi.org/10.13883/j.issn1004-5929.202404009

参考文献

基金

南省自然科学基金委员2021年科教联合课题(2021JJ60048)

PDF(1930 KB)

Accesses

Citation

Detail

段落导航
相关文章

/