基于GPGPU的数学形态学LiDAR点云快速滤波方法  

A Morphological LiDAR Points Cloud Filtering Method Based On GPGPU

作　　者：闸旋[1] 王慧[1] 程挺[2] 李鹏程[1] 陈永枫 

机构地区：[1]信息工程大学地理空间信息学院 [2]78155部队

出　　处：《测绘科学技术学报》2013年第1期73-77,共5页Journal of Geomatics Science and Technology

摘　　要：对机载激光雷达获取点云数据进行滤波时,因其数据量大,往往会耗费很多时间。在传统数学形态学LiDAR点云滤波方法基础上,提出利用基于GPU的并行技术,将大量密集运算分配到众多计算单元上并行执行,达到快速滤波的目的。通过相应的实验,验证了利用GPU并行计算技术实现的LiDAR点云快速滤波的正确性与可行性。Because of its large amount of data, filterting on the airbone LiDAR points cloud often need to spend a lot of time. On the basis of the traditional mathematical morphpolpgy LiDAR point cloud filtering, a method which adopted the parallel technique based on GPU and assigned the massive operations to be parallel executed in many computing unit to achieve the purpose of fast filtering, was proposed. Through the corresponding experiments, the correctness and feasibility of the proposed fast filtering method of LiDAR points cloud were verified.

关 键 词：激光雷达 数学形态学 点云滤波 图形处理单元 统一计算设备架构 

分 类 号：P237[天文地球—摄影测量与遥感]