摘要
从激光点云中提取建筑物信息是当前遥感数据处理的热点与难点,而构成建筑物的平面以及轮廓线的提取是LIDAR数据处理和建筑物三维建模的关键技术。本文通过分析激光点云数据中建筑物的特征,综合点云滤波、KD树索引、三维Hough变换以及Gauss球法向量统计算法的各自优点,提出了一套建筑物平面及轮廓的自动提取算法,并通过实验验证了该算法的有效性。
Building planar objects detection is the key approach for LIDAR data processing and building 3D reconstruction.In this paper,a workflow of building planar objects detection is given.Firstly,filter focus on the raw LIDAR input data using several methods,including height threshold,echo information,KD-tree clustering rules.Secondly,extended 3D Hough transform is used to detect and fit the planar objects combing the Gauss sphere normalization analysis.The experiment results show that this algorithm is excellent in auto building and planar objects detection and reconstruction in the case of basic building types.
出处
《微计算机信息》
2010年第7期205-207,共3页
Control & Automation
基金
基金申请人:李传荣
基金颁发部门:国防科技工业局
基金申请人:邸凯昌
苏国中
项目名称:深空探测中卫星与地面数据支持的探测车自发定位与制图
基金颁发部门:国家自然科学基金(40871202)
关键词
激光雷达
平面提取
KD树索引
霍夫变换
高斯球
LIDAR
building plane detection
KD-tree index
Hough 3D transform
Gauss sphere