基于主成分分析与Fisher判别的NIR木材识别  

Timber NIR Identification Based on the Principal Component Analysis and Fisher Discriminant Analysis

作　　者：杨金勇[1] 李学春 黄安民[3] 于仕兴[1] 王学顺[1] 

机构地区：[1]北京林业大学,北京100083 [2]广东省海洋工程职业技术学校 [3]中国林业科学研究院木材工业研究所

出　　处：《东北林业大学学报》2013年第12期132-134,140,共4页Journal of Northeast Forestry University

基　　金：国家自然科学基金项目(31270591)

摘　　要：以大叶桉和3类松树(马尾松、落叶松、樟子松)样品作为研究对象,利用主成分分析与Fisher判别模型结合木材近红外光谱技术进行木材树种的分类识别研究,分别建立原始光谱、一阶导数光谱和二阶导数光谱木材树种识别模型。结果显示,3组判别模型都能将桉树与松树木材树种进行有效识别,识别率为100%;而对于3类松树间的判别,经二阶导数平滑处理后,判别模型的准确率得到提高,原始光谱、一阶导数、二阶导数光谱的训练集判别模型识别率分别为86.14%,87.13%、98.02%。结果表明基于主成分分析的Fisher判别模型结合近红外光谱技术,可以对不同松树树种以及松树与桉树木材进行快速有效识别。Principal component analysis combined with Fisher discriminant analysis based on near-infrared spectroscopy was used to identify the target with Eucalyptus and three kinds of pine trees including Masson Pine,Dahurian Larch and Pinus-sylvestris. The raw spectroscopy data was pre-analyzed by first-order derivative and second-order derivative,aiming at comparing the different pre-analyze effects. Three models can identify Eucalyptus and pines much precisely. Second-order,data identification rate are higher than the raw group,and three groups of result are 86. 14%,87. 13% and 98. 02%. Therefore,the study method can be a useful tool to identify these timber samples.

关 键 词：近红外光谱 FISHER判别 主成分分析 木材识别 

分 类 号：O29[理学—应用数学]