摘要
激光雷达是目前发展迅速的一种主动遥感技术,其发射的激光脉冲能穿透树林冠层,实现森林三维结构特征的获取。为验证机载激光扫描器提取森林单木参数的可行性,该研究以海南省博鳌机场周边人工林为研究对象,使用机载激光扫描器Mapper5000(中国)获取的点云数据,探索对人工经济林单木参数估测的可行性。根据研究区的地形和林木结构特征,分别对槟榔和橡胶2个树种进行单木参数提取,使用K-means分层聚类对不同样地的林木进行单木分割,提取样地内单木树高、冠幅、胸径、材积和地上生物量。结果表明,2个树种的单木分割正检率均在85%以上,总体平均正检率在89.98%以上;单木树高、冠幅、胸径、材积、地上生物量估测结果的决定系数均达到0.8以上,说明该点云数据对提高森林参数估测精度有积极作用,机载激光雷达技术在森林资源精细调查中有较大的应用潜力,同时也可应用于相关果树生长情况监测,为数字果园的发展提供技术支撑。
Light Detection and Ranging(LiDAR)is one rapidly emerging type of active remote sensing at present.The laser pulse can partially penetrate the shelter of the forest canopy,further realizing the acquisition of three-dimensional structure characteristics for the whole forest.In this study,a systematic evaluation was made on the individual tree parameters of plantation forestry in Hainan Boao of China using the point cloud data.The pre-processing operation was carried out to implement the normalized point cloud data for the extraction of parameters.First,the outlier was used to remove the noise in the point cloud.The ground points were also separated by the Triangulated Irregular Network(TIN).Then,the Digital Elevation Model(DEM)and Digital Surface Model(DSM)were generated by the Kriging and TIN interpolation.First of all,different operations were selected to generate the Canopy Height Model(CHM).The elevation normalization was then performed on the point cloud data for subsequent segmentation and parameter extraction of the individual tree.K-means clustering was used to segment the images of the trees using different tree species,according to the actual topography and forest structure characteristics in the study area.The layer-by-layer clustering was used to extract the point cloud of the individual tree,the position of which was then compared with the measurement.The correct recognition rate and the recall rate of each sample plot were also calculated to analyze the position error of individual tree segmentation.Then,the local maximum method of the variable window was used to detect the vertex position of individual tree,where the pixel value of the tree vertex was taken as the estimated height of the individual tree.The average value of the individual tree canopy was calculated,according to the difference between the maximum and minimum of point cloud data for the individual tree in the east-west and north-south directions.Individual Diameter at Breast Height(DBH),volume,and aboveground biomass were calculated,a
作者
高凌寒
张晓丽
陈园园
Gao Linghan;Zhang Xiaoli;Chen Yuanyuan(The key Laboratory for Silviculture and Conservation of Ministry of Education,Precision Forestry Key Laboratory of Beijing,Beijing Forestry University,Beijing 100083,China)
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2021年第16期169-176,共8页
Transactions of the Chinese Society of Agricultural Engineering
基金
国家重大科学仪器设备开发专项(2013YQ12034304)。
关键词
遥感
模型
林业
机载LIDAR
单木参数
聚类算法
remote sensing
models
forestry
airborne LiDAR
individual tree parameters
clustering algorithm
