基于高程突变TIN的改进机载LiDAR点云滤波算法  

Improved TIN Filtering Based on Height Jump for LiDAR Piont Clouds Date

作　　者：邢旭东[1] 王星晨[2] 吕现福[3] 

机构地区：[1]北京航天勘察设计研究院有限公司,北京100070 [2]中铁大桥勘测设计院集团有限公司,湖北武汉430050 [3]西南大学地理科学学院,重庆400715

出　　处：《测绘与空间地理信息》2015年第12期105-107,110,共4页Geomatics & Spatial Information Technology

摘　　要：在研究高程突变TIN的滤波算法基础上,对该算法进行了改进。首先,进行粗差剔除,之后以Mean Shift算法进行点云分类,并分块构造D-TIN,分别设置动态坡度阈值;最后,对滤波结果进行定性与定量分析。实验表明:改进的算法能够有效地滤除原算法中不能滤除的植被及其他地物,提高了滤波的准确性。The paper improves the algorithm by studying the TIN filtering based on height jump for Li DAR Piont clouds date. Firstly,gross error elimination. Secondly,D- TIN is structured based on point clouds classification by Mean Shift and dynamic gradient threshold is set. Lastly,it qualitative and quantitative the filtering results. Experiments show that the improved algorithm provides high accuracy which can effectively remove the vegetation and other objects that the original algorithm cannot.

关 键 词：机载激光雷达 滤波 高程突变 不规则三角网 均值漂移 

分 类 号：P237[天文地球—摄影测量与遥感]