基于近红外光谱的橄榄油中玉米油掺杂量检测

Adulteration detection of corn oil in olive oil based on near-infrared spectroscopy

利用近红外光谱和模式识别技术建立了橄榄油中掺杂玉米油的快速鉴别方法。对191个橄榄油样本及混合油样本（玉米油和橄榄油）进行近红外光谱扫描，并对近红外光谱进行一阶导数和平滑处理，利用主成分分析法（PCA）对数据进行降维，通过前三个主成分的栽荷图确定了相关性较大的特征波段（7520-6927cm^-1,6270-5440cm^-1和4970-4400cm^-1），分别在3个波段内利用偏最小二乘回归（PLS）方法建立了玉米油含量的预测模型。结果表明在6270～5440cm^-1波段内，因子数为7，建立的模型精密度和稳定性最好，在玉米油质量分数为0．5％-100％的范围内，校正集样本和检验集样本的R：均能达到0．9999，标准偏差在0．126～0．139之间。因此，利用近红外光谱可以实现橄榄油品质的快速无损分析，以合频波段（6270～5440cm^-1）为建模区域可以得到更好的预测效果。

A rapid method was developed for quantification detection of adulteration of corn oil in olive oil based on near-infrared spectroscopy（NIRS）. A total of 191 olive oil samples and mixed oil samples（olive oil and corn oil） were analyzed by NIRS, respectively. After first derivative and smooth processing, principal component analysis （peA） was used to reduce the dimensionality of the spectral data. Through the loading graph of the first three principal components ,characteristic wave band （7520-6927cm-1 , 6270-5440cm-1 and 4970-4400cm-1, respectively） with max -relativity was determined. Predictive models of corn oil content were established in three bands using the partial least squares （PLS） method, respectively. The results showed that the model with best accuracy and stability were in 6270-5440cm-1 band and 7 factors. In the range of 0.5% to 100%, the R^2 of calibration samples and test sample were 0.9999, and the standard deviation between 0.126-0.139. The results indicated that it would be feasible to destructive analysis the content of corn oil in olive oil by NIRS, and better prediction results could be obtained in 6270-5440cm-1 band.