基于kd＿tree算法和法向量估计的点云数据精简方法

Point Cloud Data Reduction Method Based on Kd-tree Algorithm and Normal Vector Estimation

为了提高海量点云数据建模的效率,对传统的数据精简算法进行改进,提高点云数据精简精度,保留点云数据基本特征,提出基于kd_tree算法和法向量估计的点云数据精简方法。该方法利用kd_tree实现每个点K近邻数据搜素,构建点云数据树状拓扑关系,通过最小二乘拟合K近邻点平面,计算平面法向量,实现每个点云数据法向量的获取,根据点云数据法向量夹角关系,实现点云数据精简。实验证明,这种方法能够很好保留点云特征信息,实现点云数据精简。

In order to improve the efficiency of massive point cloud data modeling and promote the traditional data reduction algorithm.The point cloud data reduction method based on kd_tree and normal vector estimation is proposed.It improved the accuracy of point cloud data reduction,retained the basic characteristics of point cloud data.The method searched the K-nearest neighbor data of each point through kd_tree,built topological relationships and calculated the plane normal vector using the least squares principle.It obtained the normal vector of each point cloud data and simplified the point cloud data according to the normal vector angle relationship of the point cloud data.The experiment shows that this method can preserve the feature of point cloud and simplify the point cloud data.