基于机载LiDAR点云数据的组合滤波方法研究
Research on Combined Filtering Method Based on Airborne LiDAR Point Cloud Data
沈豫 1葛灵斌 2洪年祥3
作者信息
- 1. 舟山市自然资源测绘设计中心,浙江 舟山 316021
- 2. 台州市自然资源和规划局,浙江 台州 318000
- 3. 浙江省测绘科学技术研究院,浙江 杭州 311100
- 折叠
摘要
为提高机载LiDAR点云的滤波精度,更好地服务于高精度数字高程模型(Digital Elevation Model,DEM)制作,本文综合各种机载LiDAR点云滤波方法的优点,提出一种组合点云滤波方法.该组合滤波算法由渐进三角网加密(Progressive TIN Densification,PTD)滤波、邻域高差滤波以及平面拟合滤波算法组成.首先,使用PTD算法对原始机载LiDAR点云数据进行初始滤波,滤除大部分地物点;其次,使用邻域高差滤波算法补全缺失点云区域;最后,使用平面拟合滤波算法滤除近地噪声点.本文提出的组合滤波算法综合了上述 3 种滤波算法的优势,提高了对地形的自适应性以及近地点去噪的处理能力,实验表明本文提出的组合滤波算法能有效弥补单一PTD滤波算法难以保留陡坡区域地形特征的缺陷,有效提升点云滤波的质量.
Abstract
In order to improve the accuracy of airborne LiDAR point cloud filtering and better serve the production of high-accuracy digital elevation model(DEM),this paper integrates the advantages of various airborne LiDAR point cloud filtering methods and pro-poses a combined point cloud filtering method.The combined filtering algorithm is composed of Progressive TIN Densification(PTD)filtering,neighborhood height difference filtering and plane fitting filtering algorithms.First,the PTD algorithm is used to perform ini-tial filtering on the original airborne LiDAR point cloud data to remove most of the ground objects;secondly,the neighborhood height difference filtering algorithm is used to complete the missing point cloud region;finally,the plane fitting filtering algorithm is used to filter out the near-ground noise points.This paper proposes a combined filtering algorithm that integrates the advantages of the above three filtering algorithms,which improves the adaptability to terrain and the processing ability of near-ground denoising.The experi-ment also shows that the combined filtering algorithm in this paper can effectively improve the shortcoming that a single PTD filtering algorithm is difficult to retain the terrain features of steep slope areas,and effectively improve the quality of point cloud filtering.
关键词
机载LiDAR点云滤波/渐进三角网加密/邻域高差滤波/平面拟合滤波Key words
airborne LiDAR point cloud filtering/progressive TIN densification/neighborhood height difference filtering/plane fitting filtering引用本文复制引用
出版年
2024
NETL
NSTL
万方数据