文献类型：期刊文献

中文题名：基于近红外光谱快速鉴别木材种类的研究

英文题名：A new method based on near infrared spectra for rapid identification of timber species

作者：崔宏辉[1] 房桂干[1] 梁龙[1] 吴珽[1] 刘雯雯[1]

第一作者：崔宏辉

机构：[1]中国林业科学研究院林产化学工业研究所江苏省生物质能源与材料重点实验室国家林业局林产化学工程重点开放性实验室生物质化学利用国家工程实验室,江苏南京210042

年份：2015

卷号：0

期号：2

起止页码：169-171

中文期刊名：现代化工

外文期刊名：Modern Chemical Industry

收录：CSTPCD;;Scopus;北大核心:【北大核心2014】；CSCD:【CSCD2015_2016】；

基金：国家林业局948项目(2014-4-31)

语种：中文

中文关键词：木材;近红外漫反射光谱;主成分分析;判别分析

外文关键词：wood; near infrared diffuse reflection spectroscopy; principal component analysis; discrimination analysis

分类号：O675.3;TS721

摘要：通过采集5种木材近红外光谱特征曲线,结合化学计量学技术建立预测模型,对不同种木材进行快速无损检测.设计正交试验L24（2 ×4 ×3 ×8）对光谱特征曲线进行优化预处理.利用TQ8.6软件对正交试验结果进行判别分析.结果显示,选择多元散射校正（MSC）、Savitzky-Golay平滑对波段10000～4000 cm-1进行光谱预处理,提取主成分为8时,光谱判别正确率为100％.将22种已知种类木材光谱信息导入预测模型,进一步评估预测模型的性能,验证通过率为100％,获得了较理想的预测结果.该方法的建立为木材种类鉴别提供了一个新方法,同时为其他木种植物的鉴别提供参考.
Near infrared diffuse reflection spectra of 5 timber species combined with chemometric statistical analysis are used to build a prediction model, in order to identify different species of timber quickly and non-destructively. Orthogonal experiment L24（2 × 4 × 3 × 8 ） is designed to optimize the spectrum characteristic curve. The result of orthogonal experiment is analyzed by the discriminant analysis using TQ8.6. It reveals that when multiple scattering correction （MSC） and the Savitzky-Golay smooth are used to treat 10 000 -400 cm^-1 band, by extracting 8 kinds of principal extracting components, the spectral discriminant accuracy is 100%. The performance of prediction model is evaluated by inputting 22 kinds of known species of wood to prediction model. The test pass rate also reaches 100% and the results are satisfied. This method provides a new method for the identification of timber species, and a reference for the identification of other wood species of plants.