基于SLAM点云的杨树参数提取及生物量估算模型研究

LiDAR data extraction and biomass model of poplar based on SLAM point cloud

传统的生物量计算模型是在地面调查基础上,基于胸径和树高而建立。随着激光雷达技术的发展,树木的冠幅成为更易获取的参数,加入冠幅参数建立的树木生物量计算模型更具有实际应用价值。目前没有明确针对连云港杨树的生物量估算模型,为了基于冠幅计算连云港杨树生物量,该研究使用飞马SLAM100搭载背包平台(SLAM-BP)对4个杨树样地(16 m×50 m)进行数据采集,通过去噪、单木分割等处理提取胸径、树高和冠幅等单木因子,在分析地面实测数据和点云测量数据基础上,分析3种单木因子之间的关系,并建立包括冠幅因子在内的4种模型计算生物量,分别与异速生长模型进行比较,结果表明:(1)SLAM点云数据提取的单木因子与地面实测数据之间有很高的相关性,R 2均高于0.9,可以有效提取杨树的林木参数;(2)杨树冠幅与胸径、树高之间有明显的相关性,冠幅与胸径的相关性比冠幅与树高的相关性更强,为模型建立与分析提供了坚实基础;(3)利用冠幅建立模型计算生物量具有较好的精度,但进一步增加胸径或树高可以有效提高模型的精度,同时增加二者后模型精度最高。相关模型的建立为利用地面或者无人机点云数据计算杨树生物量提供了技术支持。

The traditional forest biomass calculation model is established based on the ground survey of tree DBH and height.With the development of LiDAR technology,the crown width of trees has become more easily obtained,and the tree biomass calculation model established by crown width involvement has more practical application value.In order to calculate the biomass of poplar in Lianyungang,the Pegasus SLAM100 backpack platform(SLAM-BP)was used to collect data from four poplar sample plots(16 m×50 m),individual tree factors such as DBH,tree height and crown width were extracted through denoising and single tree segmentation,and the relationship was analyzed between the three factors of tree based on ground measurement data and point cloud measurement data.The results showed that:(1)The single tree factor extracted from SLAM point cloud data had a high correlation with the ground measured data,with R 2 higher than 0.9,competent to extract the tree parameters of poplar;(2)there was a significant correlation between the crown width and both DBH and tree height of poplar,and the correlation between crown width and DBH was stronger than that between crown width and tree height,which provided a solid foundation for the model establishment and analysis.(3)The model by use of crown width to establish had good accuracy to calculate biomass,but further involvement of either tree DBH or height could effectively improve the model accuracy,and the accuracy could reach the highest at both participation.The establishment of the relevant model could provide technical support for the calculation of poplar biomass using ground or UAV point cloud data.