基于扫描线的车载激光雷达点云滤波方法  

A Point Cloud Filtering Method of Vehicle-Borne LIDAR Based on Scanning Beam

作　　者：杨洋[1] 张永生[1] 马一薇[1,2] 杨靖宇[1] 

机构地区：[1]信息工程大学测绘学院 [2]61512部队

出　　处：《测绘科学技术学报》2010年第3期209-212,共4页Journal of Geomatics Science and Technology

基　　金：信息工程大学测绘学院硕士学位论文创新创优基金(20090203);矿山空间信息技术国家测绘局重点实验室开放基金资助项目(KLM200904)

摘　　要：在分析车载激光点云扫描特征的基础上,提出了一种基于扫描线的车载激光雷达点云滤波方法。首先利用坡度差值将扫描线分成不同的线段集合,剔除原始点云数据中的离散点;然后对各线段集合赋予相应属性来进行分类;最后利用局部线性回归进一步对分类结果进行精化。对比实验表明:该滤波方法不但能够成功提取建筑物立面点,而且还能很好地保持建筑物立面的细节特征。On the basis of the analysis of vehicle-borne LIDAR scanning features,a point cloud filtering method based on the scanning beam was proposed.Firstly,slope difference should be used to segment the scanning beam into several line segment sets to eliminate the discrete points.Then each line segment set is classified by endowing with homologous attribute.At last,local linear regression is used to adjust the results.The contrast experiments showed that： the method proposed succeeds in extracting points of building facades,as well as preserving the detail property of building facades.

关 键 词：车载激光扫描 扫描线 分布特征 滤波 坡度差 

分 类 号：P237.3[天文地球—摄影测量与遥感]