首页> 中文期刊>中国烟草学报 >原烟在线近红外光谱模型转移研究

原烟在线近红外光谱模型转移研究

     

摘要

In order to realize transferring of raw tobacco online NIR model, two multi-channel photo diode array (PDA) detector online near infrared (NIR) spectrometers were applied as master and slave instruments. Predicting results of three methods, i.e. (1) offset transfer, (2) slope intercept method with spectra of master instrument, and (3) remodeling slave instrument model were compared. Results showed that compared with master instrument model, internal parameters of slave instrument model built by re-modeling, offset transfer and slope intercept method all had good coherence. External validation results showed that mean relative error (MRE) of offset transfer method was big. Both slope intercept and re-modeling method had similar and preferable MRE, demonstrated that transfer model by slope intercept method could replace re-modeling model for sample prediction.%为实现原烟在线近红外光谱模型的转移,利用两台多通道阵列检测器型在线近红外光谱仪作为主从仪器对烟叶烟碱指标的模型转移进行研究.将主仪器采集的光谱进行偏移量转移和斜率截距法转移以及从仪器采样重新建模等三种方法构建从仪器模型,比较三者的预测效果.结果显示:偏移量转移、斜率截距法转移、重新建模与主仪器模型相比,模型内部参数有较好的一致性;外部验证平均相对误差结果表明偏移量转移的平均相对误差较大,而斜率截距法转移和重新建模的外部平均相对误差水平非常接近,说明可利用转移模型替代重新建模对样本进行预测.

著录项

相似文献

  • 中文文献
  • 外文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号