基于密度聚类和投票判别的三维数据去噪方法  

Denoising method for scanned 3D point cloud based on density clustering and majority voting

作　　者：陶抒青 刘晓强[1] 李柏岩[1] Shen Jie 

机构地区：[1]东华大学计算机科学与技术学院,上海201620 [2]美国密歇根大学蒂尔伯恩分校计算机与信息科学系

出　　处：《计算机应用研究》2018年第2期619-623,共5页Application Research of Computers

基　　金：上海市教育委员会科研创新项目(12ZZ060)

摘　　要：介绍一种有效发现和去除三维数据噪声方法,它既能改善去除噪声的效果,又能保持特征信息。该方法是一种两阶段噪声数据处理方法。该方法首先通过密度聚类将数据分类为正常簇集合、疑似簇集合、异常簇集合,然后利用正常簇集合中的点对疑似簇集合中各点进行投票判断,最终得到一个合理的三维点云数据模型。实验结果证明,该方法能够有效去除制造类工件模型的三维点云数据中的噪声数据,同时能良好保持模型表面的特征,加快处理效率。This paper presented an effective outlier detection and removal method of denosing 3 D data, which aimed at improving denoising effects as well as reserving characteristic information. This paper adopted the two stage treatment method. Firstly, it classified the data as good clusters, suspicious clusters or outliers clusters according to density clustering. Then the good cluster point determined suspicious clusters by majority voting, which would get a reasonable 3D point cloud data model finally. The experimental results show that this method can effectively remove the noise data in the 3D point cloud data and can effectively maintain the characteristics of the model surface and speed up the processing efficiency.

关 键 词：点云数据 异常点检测 基于密度聚类 投票判别算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]