Voxel-based 3-D modeling of individual trees for estimating leaf area density using high-resolution portable scanning lidar
Voxel-based 3-D modeling of individual trees for estimating leaf area density using high-resolution portable scanning lidar
复制标题
DOI：
10.1109/tgrs.2006.881743
复制
发表时间：
2006-12-01
影响因子：
8.2
通讯作者：
Omasa, Kenji
中科院分区：
文献类型：
--
作者：
Hosoi, Fumiki;Omasa, Kenji
利用高分辨率便携式扫描激光雷达生成的基于精确体素的树木模型，提出了一种精确估算不同条件下小树(油茶和红花茶树)叶面积密度(LAD)和累积叶面积指数(LAI)剖面的方法。在这种基于体素的树冠轮廓(VCP)方法中，每棵树的树冠每一水平层的数据都是使用最佳倾斜的激光从树周围对称的方位测量点收集的。然后，这些数据被转换成基于体素的三维模型，精确地再现了这棵树，包括树冠内的树。这种精确的体素模型允许这些树叶极其密集和非随机分布的树木的LAD和LAI通过使用点样方方法直接计算每一层中的光束接触频率来计算。对叶片倾斜度和非光合作用组织的校正降低了。估计错误。光束入射天顶角接近57.5度时，在不知道实际叶片倾斜度的情况下，对叶片倾斜度提供了很好的校正。通过图像处理技术去除非光合作用组织。最好的LAD估计在最小水平层厚度时误差为17%，在最大层厚度时误差为0.7%。最佳叶面积指数估计的误差也为0.7%。
A method for accurate estimation of leaf area density (LAD) and the cumulative leaf area index (LAI) profiles of small trees (Camellia sasanqua and Deutzia crenata) under different conditions was demonstrated, which used precise voxel-based tree models produced by high-resolution portable scanning lidar. In this voxel-based canopy profiling (VCP) method, data for each horizontal layer of the canopy of each tree were collected from symmetrical azimuthal measurement points around the tree using optimally inclined laser beams. The data were then converted into a voxel-based three-dimensional model that reproduced the tree precisely, including within the canopy. This precise voxel model allowed the LAD and LAI of these trees, which have extremely dense and nonrandomly distributed foliage, to be computed by direct counting of the beam-contact frequency in each layer using a point-quadrat method. Corrections for leaf inclination and nonphotosynthetic tissues reduced the. estimation error. A beam incident zenith angle near 57.5 degrees offered a good correction for leaf inclination without knowledge of the actual leaf inclination. Nonphotosynthetic tissues were removed by image-processing techniques. The best LAD estimations showed errors of 17% at the minimum horizontal layer thickness and of 0.7% at the maximum thickness. The error of the best LAI estimations was also 0.7%.