基于红外特征光谱的初榨椰子油掺假检测技术研究

根据不同油脂的红外光谱特性,建立偏最小二乘法(PLS)和反向传递神经网络(BP-ANN)判别分析模型,进行初榨椰子油中大豆油、玉米油、葵花籽油的掺假检测分析。结果发现,PLS和BP-ANN模型均具有较好的掺假检测分析能力,其中BP-ANN模型的分析效果最佳,其对初榨椰子油中大豆油、玉米油、葵花籽油进行掺假检测的准确率均达到了99.67%以上,该方法可用来进行初榨椰子油的掺假分析,具有分辨率高、快速、简便等特点。

Adulteration Detection of Virgin Coconut Oil Based on Characteristics of Infrared Spectra

According to the infrared spectral characteristics of different oils, a partial least squares ( PLS) and a back propagation artificial neural network (BP-ANN) discriminant analysis model were established to analyze the adulteration of soybean oil, corn oil and sunflower oil in virgin coconut oil. The results showed that both PLS and BP-ANN models had a good ability of quantitative detection, and the adulteration analysis was better in the BP-ANN model than in the PLS model. According to the analysis of the BP-0ANN model, the adulteration detection accuracy for virgin coconut oil mixed with soybean oil, corn oil and sunflower oil were more than 99.7%. This method could be used for adulteration quantitative detection of virgin coconut oil, and it was of high resolution, fast and easy to operate.