基于形状特征的点云简化技术研究

Research of data point simplification based on shape character

为了提高实体反求的效率,提出一种点云简化方法。该方法通过建立点云的栅格化拓扑关系,有效地收集每个测量点的邻域点,并在邻域内建立局部坐标系。同时,拟合局部抛物面估算各个测量点的曲率值,并根据曲率变化收集形状特征点。最后,依据邻域内形状特征点的分布状况,对点云实施简化。该简化方法充分地保留了原始点云的形状特征,同时删除了大量的冗余点,具有一定的先进性,为反求工程的曲面重构提供了良好的基础。

To improve the efficiency of entity reverse building, a method was given to reduce cloud data. This method searched the nearest neighboring point of every metrical point with grid topology relationship, the local reference frame was built and a local surface was fittd in the neighborhood area to compute the curvature of every metrical point. Then, the shape characteristic point was collected based on the curvature change, the point cloud was reduced according to the distributing of the shape characteristic point in the neighborhood area. This method reserved the shape character of original point cloud and deleted many otiose points, which had some advances and offered a better base for the surface rebuild of reverse engineering.