摘要
对湿地松木材近红外光谱与X射线衍射法测定的木材结晶度之间的相关性进行分析，并结合近红外光谱分析技术的基本理论，探讨降低光谱范围和选择相关光谱信息对近红外光谱预测木材结晶度的影响。结果表明：1）降低参与建模的近红外光谱范围仍然可以得到比较理想的近红外光谱模型与预测结果，当选用2000．2500nm区域的光谱建立模型时，预测值与实测值的相关系数r达到0．943；2）当选择光谱范围更小但与木材纤维素吸收峰密切相关的光谱数据（1400～1660nm或2020～2250nm）进行建模时，模型的预测效果并未降低（r〉0．947），甚至仅采用7个光谱数据也可以得到比较理想的预测结果，预测相关系数r可达到0．930，说明采用更少的但与木材纤维素吸收峰密切相关的光谱信息，所建立的预测模型仍可得到比较理想的预测效果，这将有利于低成本、便携式近红外光谱仪的开发。
The correlation between near infrared spectroscopy （NIR） and X-ray diffraction （XRD） determined crystallinity of Slash Pine wood, and the effect of reducing and selecting spectral wavelengths on prediction ability by near infrared spectroscopy were investigated in this paper. The results showed that there are strong correlations between NIR predicted and X-ray diffraction determined crystallinity based on reducing the spectral range, the correlation coefficients are 0.943 with low RMSEP values based on the reduced spectral ranges between 2 000～ 2 500 nm. Using smaller wavelength ranges （ 1 400 ～ 1 660 nm or 2 020 ～ 2 250 nm） and/or even selecting certain wavelengths consistent with known chemical absorptions did not have a significant effect on the results of NIR prediction. The result shows that the NIR prediction model based on only several spectral variables yielded high r value at 0.930 with low RMSEP value, indicating it is important to ensure that the selected wavelengths are consistent with known chemical absorptions in NIR modeling, and it＇s feasible to use smaller and lighter portable spectrometer in field applications for tree breeding programs and for wood optimal utilization in the near future.
出处
《林业科学》
EI
CAS
CSCD
北大核心
2007年第10期95-99,共5页
Scientia Silvae Sinicae
基金
国家948项目"人工林木材的近红外材性预测及增值利用技术"(2002-45
2003-4-27)
关键词
近红外光谱
X射线衍射
木材
结晶度
相关性
near infrared spectroscopy
X-ray diffraction
wood
crystallinity
correlation
