摘要
运用实验室测量的阔叶红松林的叶片光谱数据,对吉林蛟河实验区的主要树种（红松、白桦、白牛槭、春榆、裂叶榆、蒙古栎、青楷槭、色木槭和紫椴等）的叶片进行分类研究。结果表明：实验室测量叶片光谱数据,针阔树种分类精度达到100%;所有树种分类精度为80%~100%。运用波段响应函数分别模拟多光谱传感器GEOEYE-1、RAPIDEYE和WORDVIEW2的光谱,可以有效区分针阔树种,分类精度为71.6%~100.0%;所有树种分类精度为47.3%~74.0%。
We studied the leaf classifications with blade hyperspectral data for the main tree species in Jiaohe,Jilin Province.Nine tree species were Pinus koraiensis,Betulaplatyphylla Suk,Acer mandshuricum Maxim.,Ulmus japonica,Ulmus laciniata（Trautv.） Mayr.,Quercus mongolica,Acer tegmentosum Maxim.,Acer mono Maxim.and Tilia amurensis Rupr..The classification of needle-leaved and broad-leaved tree species was perfect with an accuracy of 100%.The classification accuracy among all tree species was in 80.0%-100.0%.We resampled the spectrum into several multispectral sensors（GEOEYE-1,RAPIDEYE and WORDVIEW2） by using their band response functions,and effectively distinguished coniferous species and deciduous species with the accuracy of 71.6%-100.0%.However,the classification accuracy for all species was low with47.3%-74.0%.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2015年第3期48-55,共8页
Journal of Northeast Forestry University
基金
国家"十二五"科技支撑计划项目(2012BAC01B03)
国家自然科学基金项目(41171278)
关键词
叶片光谱
混交林
树种分类
Leaf hyper spectral
Mixed forest
Tree species classification
