基于体元结构车载点云行道树提取方法研究

Research on Vehicle Point Cloud Street Tree Extraction Method Based on Voxel Structure

充分挖掘车载激光扫描系统获取地物点云三维空间信息、回波强度信息,提出一种基于体元空间特征分析的行道树提取方法。首先完成原始数据预处理提取道路附属物点云数据,建立三维体元结构;以体元结构为基本单元计算体元单元中点云回波强度、曲率特征,后分析邻域范围内体元特征联系,构建体元邻域特征描述规则,提取行道树树干结构;在树干结构提取的基础上,确定行道树位置,建立冠层投影面积模型,进而提取冠层结构。实验结果提取显示:在复杂道路场景下,算法具有一定的稳健性,能够较为完整地提取道路两侧行道树信息。

This paper fully excavates the vehicle laser scanning system to obtain the three-dimensional spatial information and echo intensity information of ground object point cloud,and proposes a street tree extraction method based on voxel neighborhood feature analysis.Firstly,the original data preprocessing is completed to extract the point cloud data of road accessories,and the three-dimensional voxel structure is established;According to the voxel structure as the basic unit,the echo intensity and curvature characteristics of the point cloud in the voxel unit are calculated,the voxel feature relationship in the neighborhood is analyzed,the voxel neighborhood feature description rules are constructed,and the trunk structure of the street tree is extracted;Based on the extraction of trunk structure,the location of street trees is determined,the canopy projection area model is established,and then the canopy structure is extracted.The experimental results show that the algorithm is robust and can extract the street tree information on both sides of the road in complex road scenes.