近红外光谱和特征光谱的山茶油掺假鉴别方法研究

发布时间：2018-07-17 01:50

【摘要】：山茶油素有"东方橄榄油"美誉,实现掺假山茶油的鉴别具有重要实用价值,采用近红外光谱技术对掺有葵花油的山茶油进行检测。分别以1%,5%,10%为梯度制备掺假比例不同的山茶油样品,并根据掺假比例将其分为A组(0%~5%)和B组(6%~10%)共11个样品,C组(15%~40%)6个和D组(50%~100%)6个样品。将每个掺假样品充分混匀后再分为9份,依次采集其1 000~2 500nm范围的吸收光谱,共获得207条光谱曲线。每组样品的光谱数据按2∶1随机分为训练集与验证集。经去除首尾噪声后,通过主成分分析法(principal component analysis,PCA)降维,并利用前四个主成分建立了鉴别山茶油不同掺假等级的主成分-支持向量机判别模型,训练集与验证集的总体判别准确率分别达96.38%和94.20%;进一步,通过对前四个主成分的载荷系数的分析,并结合原始光谱,提取建模过程中权重较大的波长并解析其化学含义,最终确定出五个特征波长:1 212,1 705,1 826,1 905及2 148nm,以此波长重新建立近红外特征光谱山茶油掺假等级判别模型,对训练集与验证集的总体判别准确率也达到了94.20%和92.75%。研究结果表明,利用近红外光谱和特征光谱均能够较好实现山茶油掺假等级的鉴别,同时所建立的近红外特征光谱模型也为设计相应的掺假山茶油实用便携式检测仪器提供了理论基础。
[Abstract]:Camellia oil is known as "Oriental olive oil", so it is of great practical value to identify the adulterated camellia oil. Near infrared spectroscopy (NIR) is used to detect camellia oil mixed with sunflower oil. Samples of camellia oil with different adulteration ratio were prepared by gradient of 10%, and were divided into group A (0%) and group B (6 10%). There were 6 samples in group C (1540%) and 6 samples in group D (50%). Each adulterated sample was fully mixed and divided into 9 parts. The absorption spectra of each adulterated sample in the range of 1 000 ~ 2 500nm were collected in turn, and a total of 207 spectral curves were obtained. The spectral data of each group of samples were randomly divided into training set and verification set according to 2:1. After removing the head-and-tail noise, the dimensionality was reduced by principal component analysis (principal component analysis), and the support vector machine (SVM) model was established to identify different adulteration grades of camellia oil by using the first four principal components. The overall discriminant accuracy of the training set and the verification set are 96.38% and 94.20% respectively. Further, by analyzing the load coefficients of the first four principal components and combining the original spectra, the wavelengths with large weights in the modeling process are extracted and their chemical meanings are analyzed. Finally, five characteristic wavelengths (1: 1 212) 1 705 (1 8261 905) and 2148 nm were determined. The model of the adulteration grade of camellia oil from near infrared spectrum was re-established. The overall discriminant accuracy of training set and verification set was 94.20% and 92.75% respectively. The results show that the identification of camellia oil adulteration grade can be achieved by using near infrared spectrum and characteristic spectrum. At the same time, the NIR characteristic spectral model also provides a theoretical basis for the design of practical portable detection instrument for adulterated camellia oil.
【作者单位】：中国农业大学工学院现代农业装备优化设计北京市重点实验室;塔里木大学机械电气化工程学院;
【基金】：国家科技支撑计划项目(2012BAK08B04)资助
【分类号】：TS227;O657.33

【相似文献】