基于密度聚类和投票判别的三维数据去噪方法
|
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 3 D point cloud data model finally. The experimental results show that this method can effectively remove the noise data in the 3 D point cloud data and can effectively maintain the characteristics of the model surface and speed up the processing efficiency.
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 3 D point cloud data model finally. The experimental results show that this method can effectively remove the noise data in the 3 D point cloud data and can effectively maintain the characteristics of the model surface and speed up the processing efficiency.
引文
[1]王幸平.基于偏微分方程的图像去噪研究[D].南京:南京理工大学,2007.
[2]苏志勋,王小超.基于法向修正及中值滤波的点云平滑[J].计算机辅助设计与图形学学报,2010,22(11):1892-1897.
[3]Teutsch C,Trostmann E,Berndt D.A parallel point cloud clustering algorithm for subset segmentation and outlier detection[C]//Proc of SPIE.2011:509-802.
[4]史利民,王仁宏.曲线曲面拟合数据点的迭代算法[J].数学研究与评论,2006,26(4):735-750.
[5]曾清红,卢德唐.基于移动最小二乘法的曲线曲面拟合[J].图学学报,2004,25(1):84-89.
[6]吕奔,高茂庭.基于密度的孤立点检测算法改进研究[J].现代计算机,2015(17):62-67.
[7]Wang Yutao,Feng H Y.Outlier detection for scanned point clouds using majority voting[J].Computer-Aided Design,2014,62(C):55-63.
[8]Sclmiitt F.An adaptive subdivision method for surface-fitting from sampled data[J].Computer Graphics,2006,20(4):179-188.
[9]张伟平.基于M树的相似图像检索技术研究[D].沈阳:东北大学,2007.
[10]周爱武,于亚飞.K-means聚类算法的研究[J].计算机技术与发展,2011,21(2):62-65.
[11]肖宿,韩国强.基于变量分离和加权最小二乘法的图像复原[J].计算机应用研究,2012,29(4):1584-1587.
[12]Birant D,Kut A.An algorithm for clustering spatialtemporal data[J].Data&Knowledge Engineering,2007,60(1):208-221.
[13]Shen Jie,Yoon D,Shehu D,et al.Spectral moving removal of nonisolated surface outlier clusters[J].Compute-Aided Design,2010,41(4):256-267.
[14]李世林,李红军.改进的最小包围球随机增量算法[J].图学学报,2016,37(2):166-171.
[15]牛为华,孟建良,王泽,等.自适应收缩函数的Contourlet变换图像去噪方法[J].图学学报,2015,36(4):593-602.
[16]李家才.反求工程中一种数据平滑的方法[J].机械与电子,2005,3(11):21-23.
[17]张金花,吴思,胡本刚.基于地面三维激光扫描仪点云数据的去噪算法研究[J].测绘与空间地理信息,2016,39(3):76-78.
[18]罗才华,周燕.基于八叉树模型的三维点云数据预处理研究[J].福建电脑,2008,12(4):81-86.
[19]Saragih J M,Cohn J F.Deformable model fitting by regularized landmark Mean-Shift[J].International Journal of Computer Vision,2011,91(2):200-215.
[20]唐娅琴.几种图像平滑去噪方法的比较[J].西南大学学报:自然科学版,2009,31(11):125-128.
[21]Fleishman S,Cohen-Or D,Silva C T.Robust moving least-squares fitting with sharp features[J].ACM Trans on Graph,2005,24(3):544-552.
[2]苏志勋,王小超.基于法向修正及中值滤波的点云平滑[J].计算机辅助设计与图形学学报,2010,22(11):1892-1897.
[3]Teutsch C,Trostmann E,Berndt D.A parallel point cloud clustering algorithm for subset segmentation and outlier detection[C]//Proc of SPIE.2011:509-802.
[4]史利民,王仁宏.曲线曲面拟合数据点的迭代算法[J].数学研究与评论,2006,26(4):735-750.
[5]曾清红,卢德唐.基于移动最小二乘法的曲线曲面拟合[J].图学学报,2004,25(1):84-89.
[6]吕奔,高茂庭.基于密度的孤立点检测算法改进研究[J].现代计算机,2015(17):62-67.
[7]Wang Yutao,Feng H Y.Outlier detection for scanned point clouds using majority voting[J].Computer-Aided Design,2014,62(C):55-63.
[8]Sclmiitt F.An adaptive subdivision method for surface-fitting from sampled data[J].Computer Graphics,2006,20(4):179-188.
[9]张伟平.基于M树的相似图像检索技术研究[D].沈阳:东北大学,2007.
[10]周爱武,于亚飞.K-means聚类算法的研究[J].计算机技术与发展,2011,21(2):62-65.
[11]肖宿,韩国强.基于变量分离和加权最小二乘法的图像复原[J].计算机应用研究,2012,29(4):1584-1587.
[12]Birant D,Kut A.An algorithm for clustering spatialtemporal data[J].Data&Knowledge Engineering,2007,60(1):208-221.
[13]Shen Jie,Yoon D,Shehu D,et al.Spectral moving removal of nonisolated surface outlier clusters[J].Compute-Aided Design,2010,41(4):256-267.
[14]李世林,李红军.改进的最小包围球随机增量算法[J].图学学报,2016,37(2):166-171.
[15]牛为华,孟建良,王泽,等.自适应收缩函数的Contourlet变换图像去噪方法[J].图学学报,2015,36(4):593-602.
[16]李家才.反求工程中一种数据平滑的方法[J].机械与电子,2005,3(11):21-23.
[17]张金花,吴思,胡本刚.基于地面三维激光扫描仪点云数据的去噪算法研究[J].测绘与空间地理信息,2016,39(3):76-78.
[18]罗才华,周燕.基于八叉树模型的三维点云数据预处理研究[J].福建电脑,2008,12(4):81-86.
[19]Saragih J M,Cohn J F.Deformable model fitting by regularized landmark Mean-Shift[J].International Journal of Computer Vision,2011,91(2):200-215.
[20]唐娅琴.几种图像平滑去噪方法的比较[J].西南大学学报:自然科学版,2009,31(11):125-128.
[21]Fleishman S,Cohen-Or D,Silva C T.Robust moving least-squares fitting with sharp features[J].ACM Trans on Graph,2005,24(3):544-552.