应用近红外光谱分析判别芝麻油掺伪的研究

Research on detection of adulteration of edible sesame oils by near infrared spectroscopy

研究了应用近红外光谱分析技术快速、准确判别芝麻油有无掺伪的方法。主要利用近红外光谱和主成分分析结合BP人工神经网络法进行了纯芝麻油、纯大豆油、掺有大豆油的掺伪芝麻油的判别研究。试验结果表明,利用BP人工神经网络法将83个校正集样品的10个主成分数据作为BP网络输入变量,建立的三层BP人工神经网络判别模型对26个测试集样品的判别率为96.15%,表明近红外光谱分析方法对纯芝麻油、纯大豆油、掺伪芝麻油具有很好的判别分类作用,该方法能有效判别芝麻油有无掺伪大豆油。

Rapid detection of adulterated sesame oil by Near Infrared Spectroscopic（NIR） method was studied.NIR Spectra and Principal Component Analysis（PCA） were combined with Back-propagation（BP） Neutral Network to detect the pure sesame oil,pure soybean oil,and sesame oil adulterated with half of soybean oil.The experiments showed that the ten principal components in 83 Calibration Sample Sets were applied as BP inputs.The discriminate rate by using the established BP Neutral Network of 26 sample sets was 96.15%.The near infrared spectroscopic method played a good role in the detection of pure sesame oil,pure soybean oil,and sesame oil adulterated with soybean oil,and offered a effective approach to the discrimination of pure and adulterated virgin sesame oil.