傅立叶变换近红外光谱在大豆蛋白质和粗脂肪 检测中的研究

目的本文以317份不同品种的大豆为原料,开展了大豆样品粉碎粒度的蛋白质和粗脂肪含量的近红外研究,以期建立大豆品质检测方法。方法 45份大豆粉碎样品经不同的过筛处理,对剩余的272份大豆样品在最优的粉碎粒度下建模分析。结果大豆粉碎过60目建模效果最好,蛋白质和粗脂肪含量近红外检测模型的内部交叉验证决定系数r~2分别为0.959和0.939;剩余272份大豆样品蛋白质含量的近红外检测模型的内部交叉验证相关系数为0.909,粗脂肪含量的近红外检测模型的内部交叉验证相关系数为0.918,外部验证蛋白质和粗脂肪决定系数R~2分别为0.944和0.911。结论近红外光谱技术可用于大豆品质指标的检测。

Detection of soybean protein and crude fat by Fourier transform-near infrared spectroscopy

Objective To research protein and crude fat content of crushed soybean with Fourier transform-near infrared spectroscopy(FT-NIR). For this purpose,317different varieties of soybean samples were prepared. Method Forty-five soybean powder samples were treated by different sieving process, the rest of 272 soybean samples were treated by modeling analysis under optimal crushed size. Results The results showed that crushing over 60 mesh sieve crushed soybean samples had the best modeling effect, coefficient of determination in internal cross-validation of protein and crude fat content were 0.959 and 0.939. Protein and crude fat content of internal cross-validation correlation coefficientofthe rest of 272 soybean samples were 0.909 and 0.918. Meanwhile, the external validation determinationcoefficient were 0.944 and 0.911. Conclusion The results showed that FT-NIR technology can be used for soybean quality detection.