基于地面激光雷达的单树枝干几何建模研究

根据基于地面激光雷达的单树枝干几何建模方法,本文提出了一种提高重建树木几何模型精度的方法。首先,根据水平集的骨架提取方法提取目标树点云数据完整的骨架点;其次,采用最小生成树的方法对提取的骨架点构建拓扑关系;第三,根据最小二乘圆拟合法和管道模型(Pipe Theory)计算骨架点对应枝干的半径;最后,应用广义圆柱体实现单树枝干几何建模。重建得到的树木枝干几何模型在树高和胸径(DBH)参数上与真实值或实测值误差均在5%以下,且枝干最大拟合误差均小于40 mm。结果表明:本文的树木枝干几何建模方法可靠性强,且重建的树木枝干几何模型精度高,可用于森林资源管理、精准农林业等领域。

Research on Geometric Modeling of Single-tree Branch Based on Terrestrial LiDAR

This paper according to the single tree branch geometry modeling method based on terrestriallidar, a method to improve the accuracy of reconstructing the geometric model of trees is proposed. Firstly, the skeleton points of the target tree point cloud data are extracted according to the skeleton extraction method of the level set. Secondly, the topology of the extracted skeleton points is con-structed by the minimum spanning tree method. Then, according to the least square round fitting method and the pipe theory, and the radius of the corresponding branches is calculated. Finally, the generalized cylinder is used to realize the geometric modeling of single tree branches. The geometric model of the reconstructed tree branches was below 5％ in the tree height and DBH parameters, and the maximum fitting error of the branches was less than 40mm. The results show that the geometric modeling method of tree branches in this paper is highly reliable and the reconstructed tree branch geometry is of high precision and can be used in the research and appli-cation of forest resource management and precision agriculture and forestry.