摘要
由于激光雷达装置系统存在的误差会导致获得的点云数据通常带有小振幅噪声或者离群点,这些噪声可分为大尺度噪声与小尺度噪声,目前很难找到一种能够同时对两类噪声有效降噪的算法。针对点云数据噪声处理提出一种统计学滤波和引导滤波相结合降噪方法。针对大尺度噪声类采用统计学滤波方法去除,针对小尺度噪声类采用引导滤波算法去除。实验结果表明,与体素化滤波、半径滤波等去噪算法相比,该算法噪声滤除率分别提升了45.5百分点和6.7百分点。该算法在去除大尺度噪声和小尺度噪声的同时,能够较好地保留原始的点云数据。
Due to the errors of lidar system,the obtained point cloud data usually have small amplitude noise or outliers.These noises can be divided into large-scale noise and small-scale noise.At present,it is difficult to find an effective algorithm to reduce the two kinds of noise at the same time.Aimed at point cloud data noise processing,a denoising method combining statistical filtering and guided filtering is proposed.Statistical filtering method was used to remove large-scale noise,and guided filter was used to remove small-scale noise.The experimental results show that compared with voxel filtering,radius filtering and other denoising algorithms,the noise filtering rate of the proposed method is improved by 45.5 percentage points and 6.7 percentage points respectively.This algorithm can remove large-scale noise and small-scale noise,and retain the original point cloud data.
作者
霍佳欣
杨家志
Huo Jiaxin;Yang Jiazhi(School of Information Science and Engineering,Guilin University of Technology,Guilin 541006,Guangxi,China)
出处
《计算机应用与软件》
北大核心
2023年第5期248-252,287,共6页
Computer Applications and Software
基金
国家自然科学基金项目(51167004)
广西自然科学基金项目(2013GXNSFBA019250)
广西嵌入式技术与智能系统重点实验室开放(主任)基金项目(2018B-04)。
关键词
统计学滤波
引导滤波
降噪
点云数据
Statistical filtering
Guided filter
Denoising
Point cloud
