基于车载点云数据的树木提取与分析

Tree extraction and analysis based on vehicle point cloud data

本文基于高速公路高精度点云数据,首先通过点云数据的分类处理实现对树木点云数据的提取,将树木点云投影到水平面,采用DBSCAN密度聚类算法实现单根树木的提取;然后在数据密集区域存在树木树冠点云重叠的区域,本文结合树干几何特征提取树干的位置信息,计算所有点云到树干中心的欧氏距离,将所有点云归类到最近的树干进行粗分割;最后根据粗分割的树木轮廓特征确定树冠模型与树冠中心,提出了采用基于密度特征的格网竞争算法对重叠的区域进行精细分割。试验表明,本文采用的树木分割方法能够实现单棵树木精确提取。

This article bases on high-precision point cloud data of highways.Firstly,the point cloud data is extracted through the classification processing of point cloud data,and then the point cloud of the tree is projected to the horizontal plane,and the DBSCAN density clustering algorithm is used to realize the extraction of a single tree.Secondly,there is an area where the tree canopy point cloud overlaps in the data-intensive area.This paper extracts the position information of the trunk and calculates all points by combining the geometric features of the trunk The Euclidean distance froms the cloud to the center of the trunk classifies all point clouds to the nearest trunk for coarse segmentation.Finally,the crown model and crown center are determined based on the rough segmented tree contour features.A grid competition algorithm based on density features is proposed to finely segment the overlapping regions.Experiments show that the tree segmentation method used in this paper can achieve accurate extraction of a single tree.