摘要
采用偏最小二乘法(PLS)建立了油茶籽油中掺杂菜籽油和大豆油的近红外光谱定量检测模型。配制不同比例(0～100%)的油茶籽油和菜籽油、油茶籽油和大豆油混合样品共256个,采集样品在10000～4000cm-1范围内的近红外透反射光谱,模型采用交互验证和外部检验来考察所建立模型的可靠性,不需进行任何光谱预处理,所建立的PLS模型相关系数为0.9997,训练集的交叉验证均方根误差(RMSECV)为0.504,预测集的预测均方根误差(RMSEP)为0.66。应用建立的模型对未知样品进行预测,并对预测值和真实值进行比较,在掺杂油含量为2.5%～100%之间范围内准确可靠,研究结果表明,采用近红外光谱技术可以实现纯茶油中菜籽油和大豆油掺杂量检测。
Near infrared spectroscopy and partial-least-squares were combined to establish quantitative detection model of camellia oils adulteration with rapeseed oil and soybean oil.Mixed samples of camellia oils adulted by rapeseed oil and soybean oil with different proportions（0～100%）,total samples was 256,samples were scanned and their near infrared reflectance spectrum were collected in 10000～4000cm-1 region,the reliability of the model established was verified by cross-validation and external test.There was no any spectral pre-treatment,the correlation coefficient of the PLS calibration model was 0.9997,root mean square error in cross-validation（RMSECV） of training sets was 0.504,root-mean-square error value（RMSEP） of validation sets was 0.66.The model was applied to predict the unknown samples,and it was found that in the addition range of 2.5%～100% adulteration,the validation values were accurate and reliable,the results showed that quantitative detection of camellia oils adulated with rapeseed oil and soybean oil was reliable by near infrared spectroscopy.
出处
《食品工业科技》
CAS
CSCD
北大核心
2012年第3期334-336,共3页
Science and Technology of Food Industry
关键词
近红外光谱
偏最小二乘法
茶油
掺杂
菜籽油
大豆油
near infrared spectroscopy（NIRS）
partial least squares
camellia oils
adulteration
rapeseed oil
soybean oil
