• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Wood and Fiber Science >Moisture content prediction below and above fiber saturation point by partial least squares regression analysis on near infrared absorption spectra of Korean pine.
【24h】

Moisture content prediction below and above fiber saturation point by partial least squares regression analysis on near infrared absorption spectra of Korean pine.

机译：通过对红松的近红外吸收光谱进行偏最小二乘回归分析，可预测纤维饱和点上下的水分含量。

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

This study was performed to predict the surface moisture content of Korean pine (Pinus koraiensis) with low moisture content (approximately 0%) and high moisture content above the FSP using near IR spectroscopy. Near IR absorbance spectra of circular specimens were acquired at various moisture contents at 25 degrees C. To enhance the precision of the regression model, mathematical preprocessing was performed by determining the three-point moving average and Norris second derivatives. After preprocessing, partial least squares regression was carried out to establish the surface moisture content prediction model. We divided the specimens into two groups based on their moisture contents. For the first group, which possessed moisture contents less than 30%, the R2 values and root mean squared error of prediction (RMSEP) of the model were 0.96 and 1.48, respectively. For the second group, which possessed moisture contents greater than 30%, the R2 values and RMSEP of the model were 0.94 and 4.88, respectively. For all moisture contents, the R2 and RMSEP were 0.96 and 5.15, respectively. As the range of moisture contents included in the prediction model was expanded, the error of the model increased. In addition, the peak positions of the water absorption band (1440 and 1930 nm) shifted to longer wavelengths at higher moisture contents.
机译：使用近红外光谱法进行了这项研究，以预测低水分含量（约0％）和高水分含量（高于FSP）的红松（Pinus koraiensis）的表面水分含量。在25℃的不同水分含量下获取圆形样品的近IR吸收光谱。为了提高回归模型的精度，通过确定三点移动平均值和Norris二阶导数进行数学预处理。预处理后，进行偏最小二乘回归以建立表面含水量预测模型。根据样品的水分含量，我们将其分为两组。对于水分含量低于30％的第一组，该模型的R 2 值和预测均方根误差（RMSEP）分别为0.96和1.48。对于第二组，其水分含量大于30％，模型的R 2 值和RMSEP分别为0.94和4.88。对于所有水分，R 2 和RMSEP分别为0.96和5.15。随着预测模型中包含的水分含量范围的扩大，模型的误差增加。此外，在较高的水分含量下，吸水带的峰值位置（1440和1930 nm）移至更长的波长。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 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