基于PCL库三维点云树枝重建

The Point Cloud of Branch Reconstruction Based on PCL Library

随着三维扫描技术的不断提高以及国内对于开源PCL(Point Cloud Library)库不断深入的研究,三维重建技术被学者们应用到诸多领域中。比如在数字城市、逆向工程等方面有了极速的发展,而对于数字城市的模拟中,实现树枝的重建是加强城市真实感最重要的部分。文中主要在图形用户界面应用程序开发框架Qt的基础上,通过利用PCL点云库实现对三维点云树枝的三维重建的方法。首先针对树枝数据采用体素化网格方法实现下采样,减少点云数量同时保持形状特征;利用Kd树存储点云实现空间划分以及搜索操作;其次通过最小二乘法平面拟合估计方法计算出点云的法向量;最后运用PCL中的贪婪投影三角化实现树枝点云的重建。结果显示,三角化重建能够较好的保留树枝的形态特征,可以为各个需树枝重建的场景中提供良好的基础。

With the continuous improvement of 3d scanning technology and in-depth research on the open source PCL(Point Cloud Library)in China,3d reconstruction technology has been applied in many fields by scholars.For example,rapid development has been achieved in digital city,reverse engineering and other aspects.In the simulation of digital city,the reconstruction of tree branches is the most important part to enhance the sense of urban reality.In this paper,based on the application development framework Qt of GUI,PCL is used to implementation the 3d reconstruction of tree branches of 3d point cloud.Firstly,the method of voxel mesh was used to realize the lower sampling for the branch data to reduce the number of point clouds and maintain the shape characteristics;Kd tree storage point cloud is used to realize space partition and search operations;Secondly,the normal vector of the point cloud is calculated by the least square method of plane fitting estimation;Finally,the greedy projection triangulation in PCL is used to reconstruct the branch point cloud.The results show that the triangulation reconstruction can well retain the morphological characteristics of the branches,and can provide a good foundation for the scenes that need the branches reconstruction.