三维有限元分析的LIDAR点云噪声剔除算法

Noise removal algorithm of LIDAR point clouds based on three-dimensional finite-element analysis

提出一种基于三维有限元分析的LIDAR点云典型噪声剔除算法。该算法首先采用空间六面体模型对原始LIDAR点云进行有限单元剖分;其次依据基于邻接关系的推理规则进行噪声单元与非噪声单元聚类;最后进一步选择更精细剖分阈值迭代剔除低矮噪声。采用国际主流机载LIDAR系统所获取的点云数据进行相关算法对比实验,结果表明三维有限元分析剔噪算法具有更好效果。

According to analysis of the limitations of traditional algorithms, such as local points fitting and frequency domain signal analysis, a typical noise removalremoval algorithm of LIDAR point clouds based on three-dimensional finite-element analysis is proposed. Firstly, point clouds is partitioned into smaller and similar units by finite elements named space hexahedron model. And then, all of the units are classified into noise units or non-noise units with adjacency-based reasoning rules. Finally, the low noise is removed by iterative processing with finer threshold. In this approach, we did experiments with a real strip data which is obtained by an international mainstream system. The result shows that finite-element analysis has good performance in noise removal.