机载激光雷达点云滤波方法研究  

Airborne LiDAR point cloud filtering methods

作　　者：单迪 江波[2] 李胜丰[2] SHAN Di;JIANG Bo;LI Shengfeng(Guizhou Zunyi 106 Geological and Mining Company Limited,Zunyi,Guizhou 563000,China;Guizhou Provincial Geological and Mineral Exploration and Development Bureau 106 Geological Team,Zunyi,Guizhou 550004,China)

机构地区：[1]贵州遵义一0六地质矿业有限责任公司,贵州遵义563000 [2]贵州省地质矿产勘查开发局一0六地质队,贵州遵义550004

出　　处：《北京测绘》2025年第4期468-472,共5页Beijing Surveying and Mapping

摘　　要：机载激光雷达点云滤波是生成高精度数字高程模型(DEM)的重要前提,传统的单一渐进式形态学滤波在机载点云滤波中存在一定的局限性,未能将一些接近地形表面的非地面点完全去除。为此,本文在渐进式形态学滤波算法的基础上,提出了一种改进布料模拟滤波算法,用于精确识别和剔除影响DEM精度的近地面点。该改进布料模拟滤波算法的实现过程包括:首先,在渐进式形态学滤波结果的基础上,采用网格化技术将点云数据进行处理,模拟并拟合地形表面,保留地形的关键特征;其次,基于点云数据的分割结果,对相关参数进行精确调整,实现最佳的滤波效果;最后,利用真实高程与拟合后高程值之间的离差标准化值,确定了一个合理的距离阈值,提升滤波结果的准确性。为了验证本文提出的滤波方法的实际效果,选取贵州省普安县某地的机载激光雷达点云数据进行实验。实验结果显示,本文提出的滤波方法在点云滤波的精准性方面较单一滤波方法上有了显著提升,其中Ⅰ类误差、Ⅱ类误差、总误差分别降低了2.92%、1.57%、3.47%,Kappa系数提升了0.0342,表现出了较高的稳定性,具有积极的推广价值。Airborne light detection and ranging(LiDAR) point cloud filtering is a crucial prerequisite for generating highprecision digital elevation model(DEM).Traditional single progressive morphological filtering has certain limitations in airborne point cloud filtering,as it fails to completely remove some near-ground points that are close to the terrain surface.To address this,this paper proposed an improved cloth simulation filtering algorithm based on the progressive morphological filtering algorithm to accurately identify and eliminate near-ground points that affect DEM accuracy.The implementation process of the improved cloth simulation filtering algorithm is as follows:first,processing the point cloud data using gridbased technology based on the progressive morphological filtering results to simulate and fit the terrain surface,while preserving key terrain features;second,making precise adjustments to the relevant parameters based on the segmentation results of the point cloud data to achieve the optimal filtering effect;finally,using the deviation normalization value between the true elevation and the fitted elevation,a reasonable distance threshold is determined to enhance the accuracy of the filtering results.To validate the practical effectiveness of the filtering method proposed in this paper,airborne LiDAR point cloud data from a region in Pu'an County,Guizhou Province,was selected for experimentation.The experimental results show that the proposed filtering method significantly improves the accuracy of point cloud filtering compared to the single filtering method.Specifically,Class I errors,Class II errors,and total errors were reduced by 2.92%,1.57%,and 3.47%,respectively,while the Kappa coefficient increased by 0.034 2.The method demonstrated high stability and has positive potential for broader application.

关 键 词：机载点云滤波 渐进式形态学滤波 改进布料模拟滤波 组合滤波算法 自适应阈值 

分 类 号：P225[天文地球—大地测量学与测量工程]