近红外快速测定饲料原料淀粉含量

摘要

NIR spectroscopy technology was used to determine the starch content of corn,Soybean meal,DDGS,etc.Compared the modeling results of different treatments of corn,wheat,barley,sorghum samples separately,the samples which smashed to 0.5mm have better result than that without treatment.We tried to build universal model of 4 kinds of grain samples,and the R,RMSECV,RPD of pretreated and treated sample models were 0.983,0.861,5.48 and 0.988,0.738,6.48,which were better than every single model of 4 grains.Therefore,it was feasible to build universal model to analyze samples,especially in the case of lack of original samples.%利用近红外光谱技术(NIR)测定玉米、豆粕、DDGS等10种饲料原料淀粉含量.对于玉米、小麦、大麦、高粱四种原粮类样品,分别以不经任何处理和粉碎0.5 mm后进行建模比较,结果显示粉碎后效果有明显提升.探索性的尝试了4种原粮类样品建立通用模型,粉碎0,5 mm前后建模的R、RMSECV、RPD分别为0.983、0.861、5.48和0.988、0.738、6.48,优于单独建模,说明建立通用模型是可行的,特别是在建模样品量较少的情况下,可以起到较好的作用.