多尺度点云噪声检测的密度分析法

当前机载激光雷达数据和影像匹配得到的点云是密集点云数据的两类主要来源,但都不可避免存在着噪声点。本文提出一种新的点云去噪算法,可适用于这两类数据中所包含的噪声点的去除。算法主要包括两步:第1步利用多尺度的密度算法去除孤立噪声和小的簇状噪声;第2步利用三角网约束将第1步中误检测为噪声的点重新归为正常点。针对真实数据进行了剔噪试验,结果表明本文提出的基于密度分析的多尺度噪声检测算法对孤立噪声和簇状噪声都有较为效,且对于质量较差的影像匹配点云的检测也能有效处理。本文算法检测率达到97%以上。

Hierarchical Outlier Detection for Point Cloud Data Using a Density Analysis Method

Laser scanning and image matching are both effective ways to get dense point cloud data , however ,outliers obtained from both ways are still inevitable .A novel hierarchical outlier detection method is proposed for the automatic outlier detection of point cloud from image matching and airborne laser scanning .There are two main steps in this method .Firstly ,the hierarchical density estimation is used to remove single and small cluster outliers .Then a progressive TIN method is used to find non‐outliers removed in the previous steps .The experimental results indicate the effectiveness of this method in dealing with the two types of points cloud data .And this method can also handle low quality point cloud data from image matching .The quantitative analysis shows that the outlier detection rate is higher than 97% .