近红外光谱法测定豆粕粗蛋白含量

Determination of crude protein in soybean meal by near infrared spectroscopy

通过近红外光谱技术建立快速检测豆粕粗蛋白含量的方法。收集240个豆粕样品为研究对象,随机分成定标集(n=222)和验证集(n=18)。利用近红外分析技术,采用10种导数预处理光谱,定标方法选用偏最小二乘法(PLS),建立了豆粕的粗蛋白预测模型。粗蛋白含量的预测决定系数(RSQ_(V))为0.922,外部验证相对分析误差(RPD_(V))为3.479,预测值与化学值之间无显著性差异(P>0.05),能够达到精准预测的效果。结果表明,近红外光谱法适用于豆粕粗蛋白含量的快速检测。

A method for rapid determination of crude protein content in soybean meal was established by near infrared spectroscopy.A total of 240 soybean meal samples were collected and randomly divided into calibration set(n=222)and verification set(n=18).The crude protein prediction model of soybean meal was established by using near infrared analysis,10 kinds of derivative pretreatment spectra and partial least squares(PLS)calibration method.The predicted coefficient of determination(RSQ_(V))of crude protein content was 0.922,the external validation relative analysis error(RPD_(V))was 3.479,and there was no significant difference between the predicted value and the chemical value(P>0.05),which could achieve the effect of accurate prediction.The results showed that near infrared spectroscopy was suitable for the rapid determination of crude protein content in soybean meal.