机载激光雷达布料模拟滤波改进算法  

Improved Algorithm for Cloth Simulation Filtering of Airborne Lidar

作　　者：陈能辉 王凯 CHEN Nenghui;WANG Kai(Jiangxi Provincial Architectural Design and Research Institute Group Co.,Ltd.,Nanchang 330046,China)

机构地区：[1]江西省建筑设计研究总院集团有限公司,江西南昌330046

出　　处：《江西测绘》2024年第3期4-7,共4页JIANGXI CEHUI

摘　　要：针对布料模拟滤波算法(CSF)在综合地形复杂区域处理超大型低矮建筑物布料粒子时,可能会粘在屋顶或低矮植被上,从而易将某些地物点错误归类为地面点的问题,论文提出结合渐进形态学的布料模拟滤波改进算法。论文选取5组实验样本进行滤波实验。结果表明,相较于单一使用CSF算法,论文所提改进算法能够有效降低城市等综合复杂地形区域的I类误差和总误差,在滤除地物点的同时能够有效地保留地形的细节特征。In cloth simulation filtering algorithm(CSF),the cloth particles may stick to the roof or low vegetation in superlarge low-rise buildings in complex terrain areas,and it is easy to classify some feature objects as ground points.In this paper,an improved cloth simulation filtering algorithm combined with progressive morphology is proposed.Five groups of samples are selected for filtering experiment.Compared with only using CSF algorithm,the improved algorithm can effectively reduce class I error and total error in complex terrain areas such as cities,and can effectively preserve the terrain details while filtering out the ground features.

关 键 词：机载LIDAR 点云滤波 渐进形态学 布料模拟 

分 类 号：P237[天文地球—摄影测量与遥感]