基于近红外建立荞麦营养成分快速检测模型

Establishment of Rapid Detection Model of Buckwheat Nutritional Components Based on Near Infrared Spectroscopy

为探索一种快速、高效测定荞麦营养成分含量的方法,从内蒙古等地收集荞麦样品66份,采集样品近红外漫反射光谱图,并参照国标法测定样品水分、灰分、脂肪、蛋白质、淀粉含量,利用样品近红外光谱指纹结合化学计量学方法建立荞麦各营养成分的快速检测模型。结果,在10000~4000 cm-1波数范围内,分别对近红外光谱进行多元散射校正+一阶导数处理、一阶导数处理+标准正态变换及去趋势算法、无预处理、二阶导数+标准正态变换及去趋势算法,结合化学测定值建立的营养成分快速检测模型的校正和预测效果最佳;所建立的水分、灰分、脂肪、蛋白质、淀粉含量快速检测模型交叉验证决定系数R 2分别为81.6010%、94.0862%、80.9423%、99.8975%、99.8576%,外部验证决定系数R 2分别为81.60%、94.09%、80.94%、99.97%、99.86%,且验证结果预测值及化学值差异不显著。建立的模型可以满足荞麦营养成分的快速检测。

In order to explore a fast and efficient method for the determination of nutrient content of buckwheat.66 buckwheat samples were collected from Inner Mongolia and other places.Near infrared spectroscopy of samples were scanned.Determine the water,ash,fat,protein and starch content reference to the national standard.Rapid detection model of buckwheat water,ash,fat,protein and starch were established by right of using chemometrics.According to the result,in the range of 10000~4000cm-1,the best way to preprocess was multiplicative scatter correction with first derivative,first derivative with standard normal varied transformation and detrending,none,second derivative with standard normal varied transformation and detrending respectively.Under these conditions,the model works best,the cross validation decision coefficient R 2 of water,ash,fat,protein and starch by near infrared quantitative model were 81.6041,94.0862,80.9423,99.8975 and 99.8576.The external validation decision coefficients were 81.60,94.09,80.94,99.97 and 99.86 respectively.There was no significant difference between the predicted and chemical values.The established model could meet the rapid detection of buckwheat nutritional components.