摘要
机载激光雷达技术已经成为快速获取城市建筑三维数字模型的有效手段,而建筑物屋顶点云提取则是建筑物三维数字模型重建的关键.为有效剔除植被和墙面点云,以及消除地形起伏对建筑点云提取精度的影响,提出一种层进式屋顶点云提取方法.首先对Li DAR点云进行滤波,在此基础上利用点云回波特性和点云法向量检测并删除非地面点中特征明显的植被点和建筑物墙面点,然后利用连通成分分析法对非地面点聚类得到初始建筑点,最后结合DTM并利用建筑物面积和高度信息分离得到建筑物屋顶点云.试验结果表明,本方法能有效地从机载点云数据中快速提取建筑屋顶点云,有效率可达85%以上.
Airborne Li DAR technology is an effective method for obtaining three-dimensional spatial data of buildings with high point density and accuracy,and building roof point cloud extraction is a key process in three-dimensional digital reconstruction. This study proposes a gradually advanced method to extract building roof points. Firstly,most of vegetation points and building wall points are removed by using multiple echo and normal characteristics,which are based on filtering results.Secondly,the remaining points are segmented into clusters by using Euclidean clustering method.Finally,by combining with a digital terrain model（ DTM）,the points on building roofs are isolated using the building area and height characteristics. The results show that the proposed method extracts roof points based on raw Li DAR data efficiently.
出处
《中国科学院大学学报（中英文）》
CSCD
北大核心
2016年第4期537-541,共5页
Journal of University of Chinese Academy of Sciences
基金
国家科技部重大科学仪器研制专项(2013YQ120343)
国家重点基础研究发展计划(973计划)(2013CB733405)资助
关键词
机载LIDAR
点云
滤波
建筑物
屋顶提取
airborne LiDAR
point cloud
filter
building
roof point extraction
