|
基于三维荧光光谱的花生油掺伪快速检测研究 |
Rapid adulteration detection of peanut oil using three-dimensional fluorescence spectroscopy |
|
DOI: |
中文关键词: 花生油 三维荧光光谱 Zernike图像矩 Xgboost算法 广义回归神经网络算法 |
英文关键词:peanut oil three-dimensional fluorescence spectroscopy Zernike moment Xgboost algorithm generalized regression neural network algorithm |
基金项目:西北农林科技大学大学生创新创业训练项目(S201910712199) |
|
Hits: 1479 |
Download times: 505 |
中文摘要: |
建立一种基于三维荧光光谱的花生油掺伪检测方法。以纯花生油和掺伪4种常见植物油的花生油为研究对象,将三维荧光光谱图处理转化为灰度图,利用Zernike图像矩直接提取三维荧光光谱灰度图的特征信息,得到的特征信息数据通过Xgboost算法和广义回归神经网络(GRNN)算法分别建立定性和定量掺伪判别模型并对其进行验证。结果表明:Xgboost算法可以有效地对掺伪的花生油进行鉴别,并准确解析其掺伪具体成分;GRNN算法可定量预测花生油掺伪含量,各检出限分别为掺伪大豆油0.2%、掺伪菜籽油1.5%、掺伪玉米油1.0%、掺伪葵花籽油0.5%。因此,该方法可对花生油掺伪进行定性和定量分析,具有快速、简便、灵敏度高等优点。 |
英文摘要: |
The adulteration detection method of peanut oil was established based on three-dimensional fluorescence spectroscopy. Using pure peanut oil and peanut oil aduterated with 4 kinds of common edible vegetable oils as the research objects, their three dimensional fluorescence spectra were processed into grayscale, and the feature information of the three dimensional spectrum grayscale was extracted directly using Zernike moment image, and then these feature information data was used to establish qualitative and quantitative discriminant models by Xgboost algorithm and generalized regression neural network (GRNN) algorithm respectively, and the models were verified. The results showed that Xgboost algorithm could identify adulterated peanut oil effectively and analyze its adulterated components accurately. GRNN algorithm could quantitatively predict the adulteration content of peanut oil, and the limits of detection were respectively 0.2% for soybean oil, 1.5% for rapeseed oil, 1.0% for corn oil and 0.5% for sunflower seed oil. In conclusion, the proposed method could realize the qualitative and quantitative analysis of adulteration of peanut oil, and had the advantages of fastness, simpleness and high sensitivity. |
查看全文 View/Add Comment Download reader |
Close |