引用本文:[点击复制]
[点击复制]
【打印本页】【下载PDF全文】 查看/发表评论下载PDF阅读器关闭

←前一篇|后一篇→

过刊浏览    高级检索

本文已被:浏览次   下载 本文二维码信息
码上扫一扫!
基于Rapideye影像的林木地上生物量估测
吴平,黄庆丰,唐雪海,陆宁辛
0
(安徽农业大学林学与园林学院,合肥 230036)
摘要:
以石台县为研究地,结合Rapideye高分遥感影像和不同森林类型样地林木地上生物量调查数据,采用Pearson双变量相关分析方法筛选模型变量,分别用多元线性回归和随机森林算法建立不同森林类型的遥感地上生物量估测模型,并进行模型估测精度对比分析。结果表明,叶绿素红边模型(CRM)与叶绿素绿波模型(CGM)2个指数与针叶林、阔叶林生物量在0.01水平上的相关性极显著,且在其多元线性回归模型和随机森林模型中两者均被挑选为建模变量。另外,与生物量相关性较强的纹理特征主要集中的红光波段和红边波段,且仅MEAN、VAR、SM3个滤波对生物量估测贡献较大,可作为建模变量。阔叶林、针叶林和针阔混交林3种森林类型的地上生物量模型估测精度均表现为随机森林模型优于多元线性回归模型。随机森林模型生物估测绝对均方误差在12.8760~36.5363之间,相对均方误差在20.20%~45.95%之间;多元线性回归生物量估测绝对均方误差在22.0425~46.4494之间,相对均方误差在34.58%~58.42%之间。
关键词:  林木地上生物量  Rapideye影像  随机森林  逐步线性回归
DOI:10.13610/j.cnki.1672-352x.20161205.029
投稿时间:2016-03-02
基金项目:中德合作BMBF-Lin2Value-033L049C; Lin2Value-CAFYBB2012013资助。
Estimation of aboveground tree biomass based on Rapideye Imagery
WU Ping,HUANG Qingfeng,TANG Xuehai,LU Ningxin
(School of Forestry and Landscape Architecture, Anhui Agricultural University, Hefei 230036)
Abstract:
Based on Pearson bivariate correlation analysis for model variables selection together with the analysis of data from the Rapideye image and sample plot aboveground biomass survey of different forest types in Shitai County, two remote sensing aboveground forest biomass estimation models of different forest types were built using the multiple linear regression and random forest algorithm method, respectively. The estimation precision of the two models was compared. The results showed that both CRM and CGM had significant correlation with the biomass of the coniferous forest and broad-leaved forest at the 0.01 level, indicating that CRM and CGM can be used for forest biomass estimation. In addition, the texture measurements which had strong correlation with biomass were mainly contained in the band of red or red edge and only MEAN, VAR and SM filter greatly contributed to biomass estimation, which can be selected to building model variables. The forest aboveground biomass estimation precision of broad-leaved forest, coniferous forest and coniferous and broad-leaved mixed forest using the random forest biomass model was better than the multiple linear regression model. The absolute mean square error of the random forest biomass estimation was between 12.8760 and 36.5363, while the relative mean square error was between 20.20% and 45.95%. The absolute mean square error of multiple linear regression biomass estimation was between 22.0425 and 46.4494, while the relative mean square error was between 34.58% and 58.42%.
Key words:  forest biomass  Rapideye Image  vegetation indices  random forest  stepwise linear regression

用微信扫一扫

用微信扫一扫