基于TLS点云的单木精细化重建方法研究
3D fine reconstruction of single tree based on TLS data
李忠凯 1王成 1习晓环 2张合兵 3王宏涛3
作者信息
- 1. 河南理工大学测绘与国土信息工程学院,河南 焦作 454003;中国科学院空天信息创新研究院数字地球重点实验室,北京 100094;可持续发展大数据国际研究中心,北京 100094
- 2. 中国科学院空天信息创新研究院数字地球重点实验室,北京 100094;可持续发展大数据国际研究中心,北京 100094
- 3. 河南理工大学测绘与国土信息工程学院,河南 焦作 454003
- 折叠
摘要
为提升单木模型的重建精度,该文提出一种基于点云连通性的单木精细化重建方法.首先,利用最短路径结合点云连通关系生成单木的初始骨架;然后,根据相邻点的位置关系优化骨架;最后,使用圆柱拟合与树叶模拟的方法,实现高精度的单木模型重建.采用不同树种、不同密度的点云数据进行验证,结果表明,本文方法模型重建的平均绝对偏差小于5 cm、均方根误差小于0.15 cm;模型的树高、胸径与真实值相比R2达0.99以上;与Adtree方法相比,平均绝对偏差减少55.14％、均方根误差下降23.95％,模型数据量降低49.51％,表明本方法重建的单木模型具有更高的准确性和稳健性.
Abstract
To enhance the reconstruction accuracy of individual tree models,a refined individual tree reconstruction method based on point cloud connectivity is proposed.Firstly,the initial skeleton of an individual tree is generated using the shortest path algorithm combined with the point cloud connectivity relationship.Then,the skeleton is optimized based on the positional relationship between adjacent points.Finally,high-precision individual tree model reconstruction is achieved through cylinder fitting and leaf simulation.The method is validated using point cloud data of different tree species and densities.The results show that the average absolute deviation of the model reconstruction in this method is less than 5 cm,and the root mean square error is less than 0.15 cm.The R2 values of tree height and diameter at breast height compared with the true values are above 0.99.Compared with the Adtree method,the average absolute deviation is reduced by 55.14％,the root mean square error is decreased by 23.95％,and the model data volume is reduced by 49.51％,indicating that the individual tree models reconstructed by this method exhibit higher accuracy and robustness.
关键词
地基激光雷达/点云/骨架提取/单木重建/点云连通性Key words
terrestrial LiDAR/point cloud/skeleton cxtraction/tree reconstruction/point cloud connectivity引用本文复制引用
出版年
2024
NETL
NSTL
万方数据