大麦、小麦和豌豆水分近红外快速分析模型的建立

Establishment of NIRS Rapid Analysis Models of Water Content in Barley,Wheat and Peas

为实现曲粮大麦、小麦以及豌豆水分的快速分析,提高样品日常的检测效率。采用近红外光谱技术对曲粮样品进行光谱扫描,并结合传统分析方法,最终分别建立了大麦、小麦以及豌豆水分指标的快速分析模型。结果表明,所建模型的线性关系均较好,R~2分别为0.9900、0.9942、0.9952,且各模型内部参数RMSEC和RMSECV值均较小,分别为0.226、0.196、0.142和0.250、0.209、0.156,各模型质量均较好;另取验证样品对各模型的预测能力进行验证,通过与传统分析方法的结果进行比较,各模型预测的平均相对误差均小于2.0%,说明所建曲粮模型均具有较好的预测能力,可用于样品日常分析检测。

In order to quickly detect the water content in barley,wheat and peas,near infrared spectroscopy(NIRS)combined with conventional analytical methods was used to establish the rapid analysis models of water content in barley,wheat and peas.The results showed that the linear relationship of all the models was good,the correlation coefficients(R2)were 0.9900,0.9942 and 0.9952,respectively,and the values of RMSEC(0.226,0.196,0.142)and RMSECV(0.250,0.209,0.156)were relatively small,which indicated that the quality of the models was good.Another samples were taken to verify the prediction ability of the models,and the results were compared with those of the conventional analytical methods.The average relative errors of all the models were less than 2.0%,which indicated that the models had good prediction ability,and could be used for daily production analysis.