基于随机森林回归的不同龄组思茅松人工林生物量遥感估测

以云南省景谷县思茅松人工林为研究对象,以景谷县实测思茅松单木生物量数据建立其单木生物量模型,计算得出90个景谷县思茅松实测样地林分单位面积生物量,采用2005年景谷县TM遥感影像提取9个植被指数作为备选自变量,基于随机森林回归建立总体样本及各龄组样本思茅松单位面积生物量估测模型。以像元为单位,利用估测模型,并采用2005年森林资源二类调查小班数据估算景谷县思茅松人工林的生物量。结果表明：各模型的决定系数（R2）〉0.89,均方根误差（RMSE）〈7.00,预估精度（P）〉87.00%;研究区思茅松人工林单位面积生物量为59.0889 t/hm^2,其中幼龄林为38.5170 t/hm^2,中龄...

Full description

Saved in:
Bibliographic Details
Published in广东农业科学 Vol. 42; no. 15; pp. 148 - 153
Main Author 孙雪莲 舒清态 欧光龙 张博 胥辉
Format Journal Article
LanguageChinese
Published 西南林业大学林学院,云南昆明,650224 2015
Subjects
思茅松
生物量
随机森林回归
随机森林回归
生物量
思茅松
Online AccessGet full text
ISSN1004-874X

Cover

More Information
Summary:以云南省景谷县思茅松人工林为研究对象,以景谷县实测思茅松单木生物量数据建立其单木生物量模型,计算得出90个景谷县思茅松实测样地林分单位面积生物量,采用2005年景谷县TM遥感影像提取9个植被指数作为备选自变量,基于随机森林回归建立总体样本及各龄组样本思茅松单位面积生物量估测模型。以像元为单位,利用估测模型,并采用2005年森林资源二类调查小班数据估算景谷县思茅松人工林的生物量。结果表明：各模型的决定系数（R2）〉0.89,均方根误差（RMSE）〈7.00,预估精度（P）〉87.00%;研究区思茅松人工林单位面积生物量为59.0889 t/hm^2,其中幼龄林为38.5170 t/hm^2,中龄林为53.6626 t/hm^2,近熟林为94.8018 t/hm^2。
Bibliography:44-1267/S
Pinus kesiya var. langbianensis; biomass ; random forest regression
SUN Xue-lian, SHU Qing-tai, OU Guang-long, ZHANG Bo, XU Hui （ College of Forestry, Southwest Forestry University, Kunming 650224, China ）
Taking the Simao pine （ Pinus kesiya var. langbianensis ） plantations in Jinggu county as the research object, this paper constructed the biomass model of single tree and calculated the biomass of Simao pine＇s per unit of 90 sample plots of Jinggu county. Nine indexes of vegetation were selected as the alternative variables using TM remote sensing image of Jinggu country in 2005. The estimation model of Simao pine＇s per unit biomass of the overall sample and the sample of different ages was established based on random forest regression. The biomass of Simao pine＇s per unit artificial forest was predicted in study area by taking the pixel as unit in 2005. The results were as follows ： Rz 〉 0.900, RMSE 〈 7.00, prediction accuracy （ P ） 〉 87.00%. The biomass of per unit artificial forest was 59.0889 t/h
ISSN:1004-874X