摘要
杆状物作为道路场景中重要的公共设施,研究如何对其进行自动化精确分类十分重要。本文基于车载激光扫描点云数据,提出一种基于聚类的杆状物自动提取方法。主要实现步骤为:首先,对原始道路车载激光点云数据进行水平面投影并构建格网,以格网为单位进行地物点提取;其次,基于格网对地物点进行聚类;最后,以聚类结果的单个点云块为处理单元,根据地物的空间表达特征实现杆状物的精细提取与分类。为了对本文提出杆状物方法的有效性进行检验,使用实测道路点云数据进行实验,并将杆状物提取结果与人工提取结果进行对比,结果表明,灯杆与行道树均取得较高的探测率,证明了算法的正确性与优越性。
As an important public facility in road scenes,it is particularly important to study how to automatically and accurately classify pole shaped objects.This article proposes a clustering based automatic extraction method for pole shaped objects based on vehicle-borne laser scanning point cloud data.The main implementation steps are as follows:first,it projects the original road vehicle-borne laser point cloud data horizontally and constructs a grid,and extracts ground points on a grid basis;secondly,it conducts clustering of ground points based on grid;finally,it uses a single point cloud block from the clustering results as the processing unit,fine extraction and classification of pole shaped objects are achieved based on the spatial expression characteristics of the features.In order to verify the effectiveness of the proposed pole shaped object method in this article,experimental results were conducted using measured road point cloud data,and the results of pole shaped object extraction were compared with those of manual extraction.The results showed that both lamp poles and roadside trees achieved high detection rates,proving the correctness and superiority of the algorithm.
作者
黄梦霞
HUANG Mengxia(Zhejiang Institute of Surveying and Mapping Science and Technology,Hangzhou 310000,China)
出处
《测绘与空间地理信息》
2024年第8期165-167,共3页
Geomatics & Spatial Information Technology
关键词
车载激光扫描
杆状物
提取
格网划分
聚类
vehicle-borne laser scanning
pole shaped objects
extraction
grid division
clustering
