基于归一化植被点云的林分平均高及蓄积量反演  

The Inversion of Average Stand Height and Stock Volume based on Normalized Vegetation Point Cloud

作　　者：王照利 王浩伟 杨佳乐 段梦琦 马胜利 WANG Zhaoli;WANG Haowei;YANG Jiale;DUAN Mengqi;MA Shengli(Northwest Surveying,Planning and Designing Institute of National Forestry and Grassland Administration,Xi′an,710048,China;Aerial Photography and Remote Sensing Group Co.Ltd.,Xi′an 710199,China)

机构地区：[1]国家林业和草原局西北调查规划设计院,西安710048 [2]中煤航测遥感集团有限公司,西安710199

出　　处：《林业资源管理》2021年第6期37-42,共6页Forest Resources Management

基　　金：中国煤炭地质总局自然资源智能感知科技创新团队(ZMKJ-2020-T04);国家林业和草原局自主研发项目(LC-1-05);中煤总局科技项目(ZMKJ-2020-J09-02)。

摘　　要：提出一种归一化植被点云计算方法,利用植被点云与地面点云的垂直高程差表征去除地形影响的林木绝对高度值,在此基础上提取森林特征变量,使用随机森林算法对研究区林木平均高及蓄积量进行反演估测。结果表明,该方法能够有效提高森林因子的估测精度,林木平均高及蓄积量的拟合精度分别为0.946和0.936。This paper proposed a normalized vegetation point cloud computing method,which used vertical elevation difference characterization between vegetation point cloud and ground point cloud to remove the absolute height value of forest influenced by topography,on this basis,it extracted forest characteristics variables,and used the random forest algorithm to invert and estimate the average tree height and the forest stock volume within the study area.The result showed that this method can effectively improve the estimation accuracy of forest factors,and the fitting accuracy of average tree height and forest stock volume were 0.946 and 0.936,respectively.

关 键 词：激光雷达 归一化植被点云 森林因子反演 储备林 

分 类 号：S758.5[农业科学—森林经理学]