基于模糊C均值聚类的机载LiDAR点云滤波

A Filtering Method of Airborne LiDAR Data Based on Fuzzy C-Means Clustering

点云滤波是机载LiDAR数据处理的重要步骤。现有滤波算法大部分要建立点云之间的索引关系,增加了算法的复杂度;或需要对原始数据进行内插,导致原有精度损失。本文在对Li DAR点云数据的高程进行统计分析的基础上,引入模糊C均值聚类分析算法,针对大区域平坦复杂城区数据,无须建立索引或进行内插,能够快速简单地实现地面点与非地面点的分类。实验结果表明,该方法切实可行,能够较好地满足精度需求。

Filtering is an important step of airborne LiDAR data processing. Most of the existing filtering algorithms need to build the index relationship between point clouds, leading more complexity of the algorithm ; or interpolate the raw data, result in loss of the o- riginal accuracy. Based on the statistical analysis of LiDAR elevation data, this paper introduces fuzzy C -means clustering algorithm for large, flat and complex urban data, quickly and easily implements the classification of ground points and non - ground points with- out index in go interpolation. The experimental results show that this method is feasible and able to meet the accuracy requirements.