木薯品质分析的近红外光谱模型建立及其应用研究

应用近红外光谱法(NIRS)建立木薯中淀粉、水分定量分析的近红外光谱数学模型,探讨了修正偏最小二乘法(MPLS)、偏最小二乘法(PLS)以及主成分回归法(PCR)等优化处理对定标模型的影响,确定了修正偏最小二乘法(MPLS)是建立模型最适合的数学方法。并对模型预测结果的准确性进行了评价。结果表明:验证集样品的化学值和近红外预测值拟合存在较好的线性关系,相关性显著。淀粉模型预测标准偏差(Sep)为0.850,系统偏差(Bias)为-0.095,相关系数(r)为0.971。水分模型预测标准偏差(Sep)为0.075,系统偏差(Bias)为0.007,相关系数(r)为0.980。淀粉、水分定量分析的NIRS数学模型具有较高的预测准确性,可应用于木薯批量收购中的品质等分析。

Study on the establishment and application of NIR spectrum model for cassava quality analysis

Application of Near Infrared Spectral method(NIRS)establish cassava starch、water quantitative analysis of the near infrared spectrum mathematical model,this paper discusses the influence of Modified Partial Least Squares(MPLS)、Partial Least Squares(PLS)and Principal Component Regression method(PCR)for calibration model.Determine the Modified Partial Least Squares(MPLS)is building a model the most suitable mathematical method.And evaluate the results of model prediction accuracy.The results show that,validation samples of the chemical value and the prediction of near infrared fitting existence good linear relationship,significantly related.Starch model predicts standard deviation(Sep)is 0.850,the System deviation(Bias)for already deviation 0.095,correlation coefficient(r)of 0.971.Moisture prediction model(Sep)the standard deviation is 0.075,the System deviation(Bias)is 0.007,the correlation coefficient is 0.980(r).Starch,moisture to quantitative analysis of NIRS mathematical model has higher prediction accuracy and can be used in the batch purchase for the quality of cassava analysis,etc.