基于微型近红外光谱仪油菜籽粗脂肪 与粗蛋白校正模型的建立

为了验证微型近红外光谱仪的现场分析实用性,利用该光谱仪测定了油菜籽中粗脂肪与粗蛋白的含量。采集油菜籽样品的近红外反射光谱,光谱经预处理和异常样本剔除后,结合偏最小二乘法回归(PLSR)建立油菜籽的粗脂肪与粗蛋白定量分析模型。结果表明,粗脂肪的模型校正相关系数(Rc)、校正均方根误差(RMSEC)、预测相关系数(Rp)和预测均方根误差(RMSEP)分别为0.9187、1.1873、0.8162和1.3895;粗蛋白的模型校正相关系数(Rc)、校正均方根误差(RMSEC)、预测相关系数(Rp)和预测均方根误差(RMSEP)分别为0.8773、0.8153、0.8033和0.7532。验证了该光谱仪在油菜籽的粗脂肪含量和粗蛋白含量检测方面是可行的,为进一步拓展微型近红外光谱仪的应用奠定了基础。

Establishment of a calibration model for rapeseed crude fat and crude protein using a miniature near-infrared spectrometer

To verify the practicability of using a miniature near-infrared spectrometer for live analysis, the crude fat and crude protein contents in rapeseeds were determined. The near infrared reflectance spectra of the rapeseeds were collected followed by pretreatment and elimination of the abnormal samples. Partial least squares regression (PLSR) model was established for crude fat and crude protein of rapeseeds. The results showed that, for the crude fat model, the correlation coefficient (Rc) and root mean square error (RMSEC) of the calibration set and the correlation coefficient (Rp) and root mean square error of the prediction set were 0.9187, 1.1873, 0.8162 and 1.3895, respectively. For the crude protein model, the correlation coefficient (Rc) and root mean square error (RMSEC) of the calibration set and the correlation coefficient (Rp) and root mean square error of the prediction set were 0.8773, 0.8153, 0.8033 and 0.7532, respectively. These results demonstrated that detection of crude fat and crude protein in rapeseeds using the miniature near-infrared spectrometer is feasible, which built a solid foundation for the further application of the miniature near-infrared spectrometer.