摘要
海量点云数据较易获取,点云简化问题已成为众多学者的研究热点。本文提出了一种基于K-近邻拟合平面点云简化算法。通过建立KD-tree索引,寻找每个点的K-近邻,然后对K-近邻进行平面拟合剔除非特征点实现点云简化。实验结果表明本文算法简化率能达到80%以上,点云特征信息保留明显,算法适用性广,稳定性强,并且点云经简化后不损害重建结果。
Point cloud simplification is becoming the hot spot with the easier way to get massive points.An algorithm of point cloud simplification based on K-neighbors fitting plane is proposed in this paper.First,we find every point Kneighbors via establishing KD-tree index,then non-feature points are removed by fitting K-neighbors to simplify point cloud.The experimental results show that simplification rate is above 80 %,feature points are reserved well,and the algorithm can be used widely with high stability.The points after simplification do not lower the result of reconstruction.
出处
《北京测绘》
2017年第01S期86-90,共5页
Beijing Surveying and Mapping
基金
航天器高精度测量联合实验室基金资助项目(201501)
