摘要
优势树种识别是林业资源调查的重要组成部分,提高优势树种识别精度对开展森林资源调查和相关研究具有重要现实意义。采用GEE(Google Earth Engine)云平台获取霍东矿区2023年1~12月哨兵2号(Sentinel-2)时间序列影像,基于连续变化检测分类算法(continuous change detection and classification,CCDC)算法及归一化退化指数(normalized difference fraction index,NDFI)构建树种的年内生长轨迹特征,提出一种结合树种“轨迹特征+光谱特征+纹理特征”的长时序遥感影像的优势树种分层识别方法,通过设置对照组“光谱特征+纹理特征”,运用分层分类和随机森林分类算法对霍东矿区油松、辽东栎、白桦、华北落叶松、侧柏、山杨、其他杨树7种优势树种进行识别。结果表明,1)通过NDFI指数可以很好地将落叶林和常绿林区分开来;2)基于“轨迹特征+光谱特征+纹理特征”的优势树种识别效果较好,总体精度可达79.6%,Kappa系数为0.742,比对照组的总体精度高出7.3%。
The identification of dominant tree species is an important part of forestry resource surveys.Improving the ac⁃curacy of dominant tree species identification has significant practical implications for conducting forest resource surveys and related research.Using the Google Earth Engine(GEE)cloud platform,we obtained Sentinel-2 time series images for the Huodong mining area from January to December 2023.The annual growth trajectory features of dominant tree spe⁃cies were constructed based on the CCDC algorithm and the NDFI index.A dominant tree species hierarchical identifica⁃tion method combining"trajectory features+spectral features+texture features"of long-time series remote sensing im⁃ages was proposed.A control group of"spectral features+texture features"was set up,and hierarchical classification and random forest classification algorithms were used to identify 7 dominant tree species(Pinus tabuliformis,Quercus wutaishansea,Betula playphylla,Larix principis-rupprechtii,Platycladus orientalis,Populus davidiana,and poplars spp.)in the Huodong mining area.The results showed that:1)The NDFI index can effectively distinguish between de⁃ciduous forests and evergreen forests;2)The dominant tree species identification based on"trajectory features+spectral features+texture features"performed well,with an overall classification accuracy of 79.6%and a Kappa coefficient of 0.742 in the study area,which was 7.3%higher than the control group.
作者
陈丹
李晶
霍江润
马天跃
闫星光
李雨霏
CHEN Dan;LI Jing;HUO Jiangrun;MA Tianyue;YAN Xingguang;LI Yufei(College of Geoscience and Surveying Engineering,China University of Mining and Technology(Beijing),Beijing 100083,China)
出处
《森林工程》
北大核心
2025年第3期505-516,共12页
Forest Engineering
基金
国家重点研发计划项目(2022YFE0127700)。
关键词
优势树种识别
GEE
时序轨迹特征
归一化退化指数
CCDC算法
时间谐波分析
Dominant tree species identification
GEE
temporal trajectory features
normalized disturbance index
CCDC algorithm
time series harmonic analysis
