一种改进的Mean Shift点云数据滤波  

An improved Mean Shift algorithm used for point cloud data filtering

作　　者：孙正林[1] 邹峥嵘[1] 吴爱琴[1] 

机构地区：[1]中南大学,湖南长沙410083

出　　处：《测绘工程》2011年第5期57-59,共3页Engineering of Surveying and Mapping

摘　　要：点云数据滤波是三维重建质量好坏的关键,为了避免在进行滤波时花费大量计算时间建立点云间的拓扑关系,提出利用改进的Mean Shift算法直接对点云数据进行处理,使其快速移动到核密度估计函数的最大值点,从而达到降低噪声的目的。Point cloud data filter is the key to the quality of three-dimensional reconstruction.The paper proposed an improved Mean Shift algorithm to process point cloud data directly to avoid spending a lot of time to establish topological relations between point clouds during filtering which makes every point moved to the maximum of the kernel density estimate function quickly,so as to achieve the purpose of reducing noise.

关 键 词：点云 滤波 Mean SHIFT 三维重建 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]