融合主成分分析与曲率特征的点云去噪方法研究  

Research on Point Cloud Denoising Method Combining Principal Component Analysis and Curvature Characteristics

作　　者：王飞文 孙五斌 郑明丹 WANG Feiwen;SUN Wubin;ZHENG Mingdan(Zhejiang Provincial Institute of Surveying and Mapping Science and Technology,Hangzhou 311100,China)

机构地区：[1]浙江省测绘科学技术研究院,浙江杭州311100

出　　处：《测绘与空间地理信息》2024年第11期182-185,共4页Geomatics & Spatial Information Technology

摘　　要：为提高三维点云数据的应用效率与质量,针对点云数据的完整去噪尤为重要,是基于点云数据进行特征提取的重要步骤之一。本文提出了一种融合主成分分析与点云曲率特征的组合去噪方法,该组合去噪方法实现点云去噪主要分为两个步骤,首先,使用主成分分析算法进行全局去噪,剔除明显噪声;其次,根据计算点云曲率信息进行进一步去噪,剔除局部小噪声。使用实测建筑物三维点云数据进行试验,试验结果表明本文提出的组合去噪方法能够在保证建筑物整体结构信息的前提下尽可能剔除无用噪声,利用基于熵理论的定量方法检验本文方法去噪效果,结果表明本文方法去噪精度较主成分分析的去噪算法、双边滤波算法、拉普拉斯算法去噪精度更高,验证了本文方法的可行性与优越性,具有较高的推广价值。In order to improve the application efficiency and quality of 3D point cloud data,complete denoising of point cloud data is particularly important,which is one of the important steps of feature extraction based on point cloud data.In this paper,a combined denoising method combining principal component analysis and point cloud curvature characteristics is proposed.This combined denoising method is mainly divided into two steps to achieve point cloud denoising.First,principal component analysis(PCA) algorithm is used for global denoising to remove obvious noise;secondly,further denoising is carried out according to the curvature information of the calculated point cloud to eliminate local small noises.The experiment is conducted with the measured three-dimensional point cloud data of buildings.The test results show that the combined denoising method proposed in this paper can eliminate useless noises as much as possible on the premise of ensuring the overall structure information of buildings.The quantitative method based on entropy theory is used to test the denoising effect of this method.The results show that the denoising accuracy of this method is higher than that of the denoising algorithm of principal component analysis,bilateral filtering algorithm and Laplace algorithm.The feasibility and superiority of the method in this paper are verified and it has high popularization value.

关 键 词：点云去噪 主成分分析 曲率特征 信息熵 

分 类 号：P231[天文地球—摄影测量与遥感]