含噪点云预处理技术研究

Research on preprocessing technique of point clouds with noise

用三维光学测量系统进行测量时,由于周围环境、人、设备等各方面的影响,测量数据中常常会掺入噪声。针对体外飞点和离群成簇噪声分别采取基于K_近邻搜索的平均距离去噪算法和改进的基于近邻点距传播的去噪算法进行处理,取得了较好的去噪效果。针对直接测量或者多次测量拼接获取的点云存在"粗糙毛刺"和点云多层重叠的状况,采用基于MLS的拟合平面投影光顺算法进行光滑处理,去除"粗糙毛刺"和打薄重叠区域。该光顺去噪预处理算法已经成功运用到三维测量系统的点云处理模块中。

Methods improved for fairing and denoising the point clouds with noises and outliers acquired by 3D measurement systems are presented. In this paper,the noises are classified as isolated noises, outlier clusters and mixed noises. We improve different denoising and fairing algorithms for different noises, and these algorithms are applied successfully in the 3D point clouds processing system.