一种顾及地形的点云自适应坡度滤波方法  

A Point Cloud Adaptive Slope Filtering Method Considering Terrain

作　　者：丁少鹏 刘如飞[1] 蔡永宁[2] 王鹏[1] DING Shaopeng;LIU Rufei;CAI Yongning;WANG Peng(Geomatics College,Shandong University of Science and Technology, Qingdao,Shandong 266590,China;Jinan Geotechnical Investigation and Surveying Institute,Jinan 250101,China)

机构地区：[1]山东科技大学测绘科学与工程学院,山东青岛266590 [2]济南市勘察测绘研究院,济南250101

出　　处：《遥感信息》2019年第4期108-113,共6页Remote Sensing Information

基　　金：国家重大仪器设备开发专项(2013YQ120343);国家重点研发计划(2016YFB0501705);国家基础测绘科技计划(2016KJ0101);山东科技大学人才引进科研启动基金项目(2016RCJJ004);山东科技大学研究生科技创新项目(SDKDYC180311)

摘　　要：针对存在多种地形的复杂地区无法使用单一阈值进行点云滤波的问题,为提高点云滤波的自动化水平,提出了一种顾及地形的点云自适应坡度滤波方法。首先创建平面格网索引,利用高程分层统计的方法对点云进行去噪,选取地面种子点;然后进行局部坡度统计,求得局部地形坡度值,对地形进行分类;最后,根据地形确定滤波窗口大小,计算窗口内的坡度阈值进行滤波。分别使用公路和矿山数据进行实验分析。结果表明,该方法可以去除复杂地形的非地面点,保留基本的地形特征。Aiming at the problem that complex regions with multiple terrains cannot be filtered with a single threshold for point cloud filtering,apoint cloud adaptive slope filtering method taking account of terrain is proposed for improving the automation of point cloud filtering.Firstly,plane grid index is created and the point cloud is denoised by using elevation hierarchical statistics,and the ground seed points are selected.Then,through local statistical analysis of slope to determine the slope value,and the terrain is classified.Finally,after determining the size of the filtering window by the terrain,the slope threshold of each window is calculated for ground filtering.Experiments are conducted with the data of road and mine,respectively.The results show that this method can effectively remove non-ground points of complex terrain and preserve topographic features.

关 键 词：点云 坡度滤波 自适应 复杂地形 多尺度 

分 类 号：P237[天文地球—摄影测量与遥感]