摘要
在分析现存点云处理方法的特性后,通过改进三角网构网算法的算法机制,提出了一种基于空间分割的分块优先级机制的三角网表面重建算法,用于重构树冠表面,实现树冠体积的准确提取.通过可视化方法对比了多种算法的点云构网效果,以实验区选定的30棵树为研究对象,利用T-LiDAR获取树冠点云数据,通过人工方法、传统算法和本文的改进算法计算树冠体积,对这些结果进行了对比分析.分析发现:四种方法之间均显示出较好的相关性(R2>=0.831),其中所提出的改进Delaunay方法拥有理想的精度,较好稳定性和最少的耗费时间.实验结果表明,提出的算法在点云(尤其是T-LiDAR数据)树冠的体积提取中具有很大的优势.结合T-LiDAR数据还可以实现树冠表面积和生物量等树冠因子的高精度快速提取.
To improve the precision and effectiveness of crown-volume measurement and calculation,the authors have analyzed the characteristics of existing methods for processing the point cloud and have proposed a crown-surface reconstruction algorithm using a triangulated irregular network and voxel-based volumetric algorithms. This algorithm,after reconstructing the surface of the point-cloud crown,can extract the crown volume. This paper compares classic Delaunay grid-construction results with those from the proposed algorithm using a visualization method and carries out algorithm complexity analysis. These efforts have confirmed that the method presented in this paper is better than the traditional algorithm from the viewpoints of grid-construction accuracy and efficiency. This research,examined 30 trees in the study area. T-LiD AR was used to obtain point-cloud data for the crown. The classical manual dendrometric method,the point-cloud measurement method,the classical Delaunay algorithm,and the method proposed in this paper were used to calculate crown volume,and the results were compared. The four methods showed a good correlation（ R2 = 0. 831）,while the improved Delaunay method presented in this paper achieved good precision,good stability,and the least calculation time. The results of these experiments proved that the proposed algorithm has a considerable advantage in crown-volume extraction from point clouds（ especially from T-LiD AR data）. The combination of the proposed algorithm with T-LiD AR data could extract crown properties such as surface area and biomass quickly and precisely.
出处
《红外与毫米波学报》
SCIE
EI
CAS
CSCD
北大核心
2016年第2期177-183,189,共8页
Journal of Infrared and Millimeter Waves
基金
Supported by National Natural Science Foundation of China(41371001)
关键词
树冠体积
T-LiDAR
三角网
点云数据
crown volume
terrestrial LiDAR
triangular irregular net
point-cloud data
