基于点云图像与树木几何特征的林分蓄积量精准计算

The traditional stand volume calculation is based on the theory that "the cross section of the stem is a circle". This theory does not consider the geometry of each tree. The difference in the true geometry of the trees leads to the fact that the calculated stand volume based on the theory must deviate from its true value. In order to accurate calculating the stand volume, the project fully mining the geometric features of the trees contained in the point cloud images scanned by terrestrial laser scanning, and attempt to construct a series of tree volume calculation methods based on the geometric characteristics of trees. The main research contents include three aspects: (1) Constructing methods for segmenting tree growth space and extracting tree point clouds from forest stand point clouds; (2) Based on geometric features of trees and stem, constructing methods for extracting tree stem points from tree point clouds, repairing stem point cloud density and filling stem point cloud hole; (3) Based on the geometric features of the stem surface, constructing methods for stand volume calculation, including the section profile curve method, the surface model of the stem surface and the solid geometry model of the stem. This project will provide accurate methods for standing tree volume and stand volume calculation that matching with the geometric features of trees, and will enrich the development of the theory and technology of forest mensuration, and will promote the application of laser measurement technology in forestry measurement, and will provide scientific data for forest management and precise improvement of forest quality.

传统林分蓄积量计算以“树干横断面是圆”为基础理论，并未考虑每株树木的几何形状。树木真实几何形状的差异导致以此理论计算的林分蓄积量必然与其真值存在着偏差。为精确计算林分蓄积量，本项目充分挖掘地面三维激光扫描仪获取点云图像中蕴含的树木几何特征，力求构建一套基于树木几何特征的林分蓄积量计算方法。其主要研究内容涵括3个方面：(1)构建在林分点云中分割树木生长空间与提取树木点云的方法；(2)基于树木与树干几何特征构建从树木点云中提取树干点云的方法、构建树干点云密度的修复与树干点云空洞的填充方法；(3)基于树干表面几何特征构建林分蓄积量的计算方法，包括基于截面轮廓曲线、基于树干表面曲面模型与基于树干实体几何模型的计算方法。本项目将提供一套与树木几何特征相匹配的活立木材积与林分蓄积量的精确测定方法，丰富测树学理论与技术的发展，促进激光测量技术在森林调查中的应用，为森林经营与森林质量精准提升提供科学数据。

“树干横断面是圆”是单木材积计算与林分蓄积量计算的理论基础，然而这一基础理论并未考虑树木的真实几何结构，这导致依据此理论计算的树干材积与林分蓄积必然与其真实值间存在偏差。为精确计算单木材积与林分蓄积量，本项目以地面三维激光扫描仪获取的树木点云为数据，构建一套基于树木几何特征的单木材积与林分蓄积量计算方法。项目主要研究内容包括：（1）构建从林分点云中单株树木定位及单株树木点云提取的方法；（2）构建从单株树木点云提取树干点云及树干点云修复的方法；（3）构建基于树干几何结构特征的树干材积计算方法，包括基于树干断面轮廓曲线、基于树干表面模型与树干四面体实体几何模型的计算方法；（4）构建树干三维几何模型及树干直径提取方法间的比较与分析。基于上述研究内容，项目组提出一系列方法，包括从林分点云中提取单木点云，单木骨架构建，提取单木树干点云，树干点云缺失区域修复，树干断面轮廓曲线构建及树干材积计算，树干表面双三次B样条曲面模型构建及树干材积计算，树干表面三角网模型与四面体实体模型构建及树干材积计算，树干断面几何中心点优化计算，树干三维几何模型构建，模拟轮尺测量直径方法等。基于上述内容与方法，本项目研究表明：对于同一株树干点云，基于四面体实体模型的树干材积计算精度优于基于树干断面轮廓曲线与基于树干双三次B样条曲面模型的计算精度；拟合圆与拟合圆柱体提取树干直径的精度相同，优于其它基于点云的树干直径提取方法，且具有更好的鲁棒性，但提取直径值略小于围尺实测直径；提高点云预处理精度有助于提升直径的提取精度；林业调查中，使用围尺与轮尺测量直径的误差很小，验证“围尺测树直径恒等于轮尺各向测径的平均值”这一理论观点。与基于传统理论的树干材积与林分蓄积量计算方法相比，本项目提出的方法充分利用地面三维激光扫描仪获取的树干表面的几何特征，为准确计算树干材积与林分蓄积量及其它相关参数提供一套精确可行的科学方法。