A new reduction algorithm for scattered points based on local surface feature was proposed. First, a dynamic spatial index structure of scattered points was established with R*-tree. Second, the local surface reference data was obtained based on the spatial index structure. Third, the local surface reference data was approached with free-form surface, and its curvature was computed. Fourth, the reduction of scattered points was realized based on its model curvature. It is proved that this algorithm can reduce point-data effectively under the conditions that preserve the surface characteristics of scattered points.

提出一种基于局部型面特征的散乱点云精简算法，该算法采用R*-tree建立点云动态空间索引结构，基于该结构快速准确获取点云局部型面参考数据；采用自由曲面逼近该数据并估算该数据的曲率，依据曲率分布状况精简点云数据。实例证明，该算法可在保留点云型面特征的基础上，快速有效地对点云进行精简。