摘要
针对采集的原始三维点云数据存在噪声、表面不光滑不利于后期三维重建的问题,提出一种自适应密度聚类与双边滤波融合的三维点云去噪平滑方法。该方法首先对点云模型进行自适应密度聚类分析,根据聚类结果删除模型中的噪声点;然后再计算采样点的k邻域,并求得利用k邻域构造采样点所在平面的法矢,进而得到双边滤波因子,以对点云模型进行平滑。实验结果表明,该算法能有效识别并去除噪声,并对点云模型进行平滑,同时还能保持原始模型的特征信息。
The original three-dimensional point cloud data collected has the problems of noise and unsmooth surface which is not conductive to three-dimensional post-reconstruction. In view of this,this paper presents a three-dimensional point cloud denoising and smoothing method,it is based on the fusion of adaptive density clustering algorithm and bilateral filtering. First,the method applies adaptive density clustering analysis on the point cloud model,and erases the noise points in the model according to clustering result. Then,it calculates the k neighbourhood of sampling point,and calculates the normal vector of the plane where the k neighbourhood is used to construct sampling points,and further obtains the bilateral filtering factor so as to smooth the point cloud model. Experimental results show that the proposed algorithm can identify and remove noise effectively and smoothes the point cloud model. At the same time,it can well keep characteristic information of the original model.
出处
《计算机应用与软件》
CSCD
2016年第10期148-152,共5页
Computer Applications and Software
基金
国家高技术研究发展计划项目(2013AA102304)
第56批中国博士后科学基金项目(2014M562457)
关键词
点云去噪
自适应密度聚类
k邻域
双边滤波
特征保持
Point cloud denoising
Adaptive density clustering
K neighbourhood
Bilateral filtering
Feature preserving
