一种改进的Mean Shift点云数据滤波

摘要

Point cloud data filter is the key to the quality of three-dimensional reconstruction. The paperproposed an improved Mean Shift algorithm to process point cloud data directly to avoid spending a lot oftime to establish topological relations between point clouds during filtering which makes every point movedto the maximum of the kernel density estimate function quickly, so as to achieve the purpose of reducingnoise.%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.