基于无人机倾斜摄影的三维绿量估算  

Three-dimensional Green Quantity Estimation Based on UAV Oblique Photography

作　　者：朱启梦 孙金华 杨柳 王婷[1] 赵辉[2] 张磊 ZHU Qimeng;SUN Jinhua;YANG Liu;WANG Ting;ZHAO Hui;ZHANG Lei(College of Forestry,Henan Agricultural University,Zhengzhou 450002,China;Henan Academy of Forestry,Zhengzhou 450008,China)

机构地区：[1]河南农业大学林学院,郑州450002 [2]河南省林业科学研究院,郑州450008

出　　处：《河南科学》2024年第5期694-704,共11页Henan Science

基　　金：河南省重大科技专项(201300111400);河南省博士后基金(202003062);河南农业大学青年创新基金(KJCX2020A06);河南省农业科学院自助创新项目(2023ZC098);河南省科技兴林项目(YLK202203)。

摘　　要：精准、快速、高效的绿化效果评估是现代林业发展的重要要求.针对传统方法绿地覆盖率无法反映绿化空间垂直结构的缺点,提出基于切片凸包算法利用倾斜摄影点云估测三维绿量的新方法,以河南农业大学文化路校区为研究区域,通过三种方法估算不同树种及树形的三维绿量,对比不同方法的测算精度,探索最优估算方法.利用无人机进行多光谱成像和倾斜摄影成像,通过面向对象的方法在数字正射影像基础上提取林木,结合实测33个树种的冠径、冠高等信息,分别利用冠径-冠高方程法(D-H)、实景三维模型法(3D)和基于切片凸包算法的点云体积测算法(PC)进行三维绿量测算.结果表明,三种方法测算值与实测值拟合的回归方程R2分别为0.84、0.92、0.94,RMSE分别为168.85、90.74、78.46 m^(3),精度分别为62.67%、87.26%、90.19%;优势树种悬铃木三维绿量最大,龙爪槐值最小;PC法测算水杉和球形树的相对误差最小,分别为0.56%和1.27%.利用倾斜摄影点云的切片凸包算法测算三维绿量具有可行性且最精准,为筛选不同树种及树形的三维绿量估算方法的选择上提供科学依据.Accurate,rapid,and efficient evaluation of greening effects is crucial for the advancement of modern forestry.Due to the limitation of the traditional method,which fails to reflect the vertical structure of green spaces,a new method based on the slice convex hull algorithm is proposed to estimate the three-dimensional green volume using oblique photography point cloud data.In this study,conducted at the Cultural Road campus of Henan Agricultural University,three methods were employed to estimate the three-dimensional green volume of various tree species and shapes.The measurement accuracy of these methods was compared to determine the optimal estimation approach.Multi-spectral and oblique photography imaging using UAVs were employed,and tree extraction was conducted using digital orthophoto by an object-oriented method.Using measured crown diameter and crown height information of 33 tree species,the three-dimensional green volume was calculated using the crown diameter-crown height equation method(D-H),the real three-dimensional model method(3D),and the point cloud volume calculation method(PC)based on the slice convex hull algorithm.The R2 values of the regression equations fitted to the measured values using the three methods were 0.84,0.92,and 0.94,respectively.The RMSEs were 168.85 m^(3),90.74 m^(3),and 78.46 m^(3),with corresponding accuracy rates of 62.67%,87.26%,and 90.19%,respectively.The largest three-dimensional green volume was found in Platanus,the dominant tree species,while the smallest was observed in Sophora japonica.The PC method yielded the smallest relative errors for Metasequoia glyptostroboides and Spherical trees,which were 0.56%and 1.27%,respectively.Utilizing the slice convex hull algorithm on oblique photography point cloud data to calculate the three-dimensional green volume is not only feasible but also the most accurate approach,and provides a scientific basis for the selection of 3D greenness estimation methods for screening different tree species and tree shapes a scientific basis fo

关 键 词：三维绿量 无人机倾斜摄影 点云 切片凸包算法 

分 类 号：P231[天文地球—摄影测量与遥感]