点云模型的双边滤波去噪算法  

BILATERAL FILTERING DENOISING ALGORITHM FOR POINT-CLOUD MODEL

作　　者：杜小燕[1] 姜晓峰[1] 郝传刚[1] 王玉梅[1] 

机构地区：[1]苏州大学计算机科学与技术学院,江苏苏州215006

出　　处：《计算机应用与软件》2010年第7期245-246,264,共3页Computer Applications and Software

摘　　要：提出一种特征保持点云模型光顺去噪算法。该算法首先为每个采样点建立k-邻域并估算法矢,然后以该点处的微切平面为视平面,设计一个双边滤波器。算法充分考虑了模型表面法向的变化和邻域点间的距离同噪声之间的关系,通过调整采样点在法向的位置,来实现一个局部的光顺去噪过程。实验结果表明算法简单高效,同时较好地保持了模型的特征。In this paper,we present a feature-preserved point-cloud model fairing denoising algorithm.First,we build k-neighbourhood for each sampling point and evaluate its normal vector.Then,we regard the tangent plant as view plant,design a bilateral filter.The algorithm thoroughly considers the variation of normal direction on model surface and the relationship of noise and distance between neighbourhood points,by adjusting the position of sampling point in normal direction to realise a local fairing denoising process.Experimental result proves that the algorithm in this paper is simple and efficient,and preserves fairly good model feature at the same time.

关 键 词：点云 双边滤波器 去噪 特征保持 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]