摘要
为探究基于背包激光雷达(Backpack Laser Scanning,BLS)点云数据提取天然林蒙古栎单木因子的精度,确定基于BLS点云数据构建胸径-树高模型的可行性,以3块天然林蒙古栎样地(30 m×30 m)的BLS点云数据为数据源,对点云数据进行预处理、归一化和单木分割等操作,并从分割后的单木中提取胸径和树高因子。通过对比提取数据与实测数据的相关性来分析单木因子的提取精度。最后利用提取出的单木因子构建天然林蒙古栎胸径-树高模型。结果表明,蒙古栎胸径提取结果的决定系数(R^(2))为0.930~0.957,均方根误差(RMSE)为0.697~0.897 cm;树高提取结果的R^(2)为0.925~0.951,RMSE为1.479~1.683 m。基于提取数据构建的天然林蒙古栎最优胸径-树高模型为模型1(抛物线模型),模型检验指标R^(2)为0.627,RMSE为2.044 m。由此得出,BLS有一定能力在天然林中进行数据采集,提取的蒙古栎胸径和树高与实测数据具有很高的相关性,结果可靠。基于BLS建立的蒙古栎胸径-树高模型精度较高,为BLS应用于天然林中的单木生长建模提供了进一步的理论基础。
To explore the accuracy of extracting individual tree factors of Quercus mongolica in natural forests based on Backpack Laser Scanning(BLS)point cloud data,and to determine the feasibility of constructing a diameter-height curve model based on BLS point cloud data,in this study,BLS point cloud data from three natural forest Quercus mongolica sample plots(30 m×30 m)were used as the data source.Then,the point cloud data were preprocessed,normalized and individual tree segmented,and the DBH and tree height factors were extracted from the segmented individual trees.The extraction accuracy of individual tree factors was analyzed by comparing the correlation between the extracted data and the measured data.Finally,natural forest Quercus mongolica diameter-height curve model was constructed by the extracted individual tree factors.The results showed that the coefficient of determination(R^(2))of Quercus mongolica DBH extraction results ranged 0.930-0.957,and the root mean square error(RMSE)ranged 0.697-0.897 cm,and R^(2)of tree height extraction results ranging 0.925-0.951 and RMSE ranging 1.479-1.683m.The optimal diameter-height curve model for Quercus mongolica in natural forests constructed based on extracted data was model I(parabolic model)with a model test index R^(2)of 0.627 and RMSE of 2.044 m.It was concluded that BLS had the capability to conduct data collection in natural forests,and the extracted Quercus mongolica DBH and tree height had a high correlation with the measured data,and the results were reliable.BLS-based Quercus mongolica diameter-height curve models were developed with high accuracy,which provided a further theoretical basis for the application of BLS to individual tree growth modeling in natural forests.
作者
杨军
王帆
宋仲禹
YANG Jun;WANG Fan;SONG Zhongyu(Heilongjiang Forestry and Grassland Investigation and Planning Institute,Harbin 150040,China;School of Forest,Northeast Forestry University,Harbin 150040,China)
出处
《森林工程》
北大核心
2023年第5期57-64,共8页
Forest Engineering
基金
国家重点研发计划项目(2022YFD2201003-02)
黑龙江省嘉荫县生态系统各碳库碳储量本底调查及动态分析项目(HFW220100057)。
