摘要
应用近红外光谱技术分别对含滑石粉的小麦面粉样品进行快速检测,对使用不同方法预处理后的光谱采用偏最小二乘法(PLS)建立定量分析模型。同时,比较各个模型内部交互验证均方根误差(RMSECV)、交互验证预测值与真实值间的相关系数(R2)和外部均方根误差(RMSEP),选取最优模型。实验表明:使用多元散射校正预处理方法所得效果最好,应用近红外光谱在分析检测小麦面粉中滑石粉含量方面有广阔的应用前景。
A rapid method was used to determine the talc - containing wheat flour with near infrared spectrum technology. Different mathematical models were built separately by using partial least squares （ PLS ） for each pretreatment method processed data. The influence of data number on building PLS model was discussed, with internal cross-validation root mean square error（ RMSECV）, cross-validation correlation coefficient between the predicted value and the true value （R2）, and External root mean square error（ RMSEP）, ehoosed the best model. The results showed that：the result gotten from Muhiplieative scatter correction （SNV） was the best. There is widest prospects on determine the talc - containing wheat flour with NIR.
出处
《农机化研究》
北大核心
2013年第7期183-187,共5页
Journal of Agricultural Mechanization Research
基金
北京市自然科学基金项目(4073031)
