基于近红外光谱法的藜麦脂肪含量快速检测

Rapid Detection on Quinoa Fat Based on Near Infrared Spectroscopy

为了探索一种快速测定完整藜麦（Chenopodium quinoa Willd）子粒脂肪含量的方法,采集100个藜麦样品的近红外光谱,运用近红外光谱法建立数学模型并进行预测。结果表明,在10 000~4 000 cm-1波长范围内,运用一阶导数＋矢量归一化光谱方法进行预处理,结合化学方法所得数据建立藜麦粗脂肪近红外光谱定量模型,校正和预测效果最佳,所得的粗脂肪近红外定量模型的交叉验证决定系数（R2cv）为0.939 3,外部验证决定系数（R2val）为0.9235。建立的脂肪近红外光谱模型,可以用于藜麦脂肪含量的快速检测。

To explore a rapid determination method of fat content in grains of quinoa, collecting the near infrared spectra of100 quinoa samples,the predicted models for quantitative analysis of protein contents in the grains was built using near infrared transmittance spectroscopy（NITS）. Consequently,in the wavelength range of 10 000 ~4 000 cm-1,via first derivative ＋vector normalization preprocessing and combining with the data from chemical methods,we set up near infrared quantitative model of quinoa crude fat,meanwhile,calibration and prediction effect are best,and then the cross validation decision coefficient（R2cv）and external validation decision coefficient（R2val） of fat by near infrared quantitative model were 0.939 3 and0.923 5. In the end,the model of NITS about complete grains quinoa fat can be available for fast detecting quinoa fat content.