| 吴霞1，周德旭2，宗伟凯3.荧光光谱结合DM-BPNN算法快速检测
特级初榨橄榄油中掺伪植物油[J].中国油脂,2025,50(10):.[WU Xia1, ZHOU Dexu2, ZONG Weikai3.Rapid detection of extra virgin olive oil adulterated with vegetable oil by fluorescence spectroscopy combined with DM-BPNN algorithm[J].China Oils and Fats,2025,50(10):.] |
| 荧光光谱结合DM-BPNN算法快速检测
特级初榨橄榄油中掺伪植物油 |
| Rapid detection of extra virgin olive oil adulterated with vegetable oil by fluorescence spectroscopy combined with DM-BPNN algorithm |
|
出版日期：
|
| DOI：10.19902/j.cnki.zgyz.1003-7969.240383 |
| 中文关键词: 扩散映射 反向传播神经网络 激光诱导荧光光谱 特级初榨橄榄油 掺伪 |
| 英文关键词:diffusion maps back-propagation neural network laser induced fluorescence spectroscopy extra virgin olive oil adulteration |
| 基金项目:国家自然科学基金（U1904116） |
|
| Author Name | Affiliation | | WU Xia1, ZHOU Dexu2, ZONG Weikai3 | 1.College of Information Engineering, Henan Vocational College of Water Conservancy and
Environment, Zhengzhou 450008, China
2.School of Food Science and Technology, Henan
Agricultural University, Zhengzhou 450002, China
3.School of Information and
Engineering, Zhengzhou University, Zhengzhou 450001, China |
|
| 摘要点击次数: 1204 |
| 全文下载次数: 700 |
| 中文摘要: |
| 为实现特级初榨橄榄油中掺伪低价格植物油的快速、无损检测，基于扩散映射结合反向传播神经网络（DM-BPNN）算法模型实现紫外激光诱导荧光光谱检测特级初榨橄榄油中掺伪低价格植物油。在特级初榨橄榄油中掺伪不同比例的菜籽油、大豆油、葵花籽油，采用360 nm的紫外激光作为荧光光谱的激励源，通过荧光探头激发荧光光谱信号，分析了掺伪植物油和特级初榨橄榄油的荧光光谱特征和差异，解译并判定荧光光谱谱峰对应的化学物质，利用DM算法提取荧光光谱的特征信息并完成高维数据向低维数据的映射，基于BPNN算法实现对混合油中特级初榨橄榄油含量的预测。结果表明：激光诱导荧光光谱技术结合DM-BPNN算法模型可以满足对特级初榨橄榄油中掺伪菜籽油、大豆油、葵花籽油的测定，其中定量模型测试集的决定系数均大于0989 0，均方根误差均小于0.001 0。此外，3个定量分析模型可实现一定程度的交叉应用，具有很好的通用性。综上，采用荧光光谱结合DM-BPNN算法可实现对特级初榨橄榄油掺伪低价格植物油的快速检测。 |
| 英文摘要: |
| To achieve the rapid, nondestructive and noninvasive detection of extra virgin olive oil adulterated with low-cost vegetable oil, an algorithm model based on diffusion maps and back-propagation neural network(DM-BPNN) was proposed to detect of extra virgin olive oil adulterated with low-cost vegetable oil by laser induced fluorescence spectroscopy. A 360 nm UV laser was used as the excitation source of the fluorescence spectroscopy, and the laser induced fluorescence signal generated by the mixture of different proportions rapeseed oil, soybean oil, sunflower seed oil with extra virgin olive oil was excited by the fluorescence probe. The fluorescence spectral characteristics and differences between the low-cost vegetable oil and extra virgin olive oil were analyzed, and the chemical components corresponding to the fluorescence spectral peaks were interpreted and determined. The characteristic information of the fluorescence spectra was extracted using the DM algorithm, and the mapping from high-dimensional data to low-dimensional data was completed. The extra virgin olive oil content in mixed vegetable oil was predicted based on the BPNN algorithm. The results showed that laser-induced fluorescence spectroscopy combined with DM-BPNN model could meet the detection of extra virgin olive oil adulterated with low-cost vegetable oil such as rapeseed oil, soybean oil, sunflower seed oil. The determination coefficients of the quantitative model test set were more than 0.989 0, and the root mean square errors were less than 0.001 0. In addition, the three quantitative analysis models could achieve a certain degree of cross application and had good universality. In conclusion, the use of fluorescence spectroscopy combined with DM-BPNN algorithm can achieve rapid detection of extra virgin olive oil adulterated with low-cost vegetable oil. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
