论文部分内容阅读
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。
In this paper, 317 different varieties of soybeans were used as raw materials to study the near-infrared (IR) of protein and crude fat content of soybean samples in order to establish a soybean quality testing method. Methods Forty-five soybean samples were sieved and the remaining 272 soybean samples were modeled at the optimal particle size. Results Soybean had the best effect of 60-mesh modeling. The internal cross-validated determination coefficients of protein and crude fat content in NIR models were 0.959 and 0.939, respectively. The contents of the remaining 272 soybean samples were determined by NIR detection model The correlation coefficient of cross validation was 0.909, the correlation coefficient of internal cross validation of crude fat content near infrared detection model was 0.918, and the external validation protein and crude fat determination coefficient R ~ 2 were 0.944 and 0.911 respectively. Conclusion Near-infrared spectroscopy can be used to detect the quality of soybean.