本文为英国伯恩茅斯大学(作者:Navpreet Kaur Pawar)的硕士论文,共33页。
三维采集技术的最新进展为无组织三维点数据集提供了巨大的输入。随着使用点云的应用领域的增加,对重建连续曲面的需求也越来越大,从而能够提供无组织点集的真实表示,并呈现用于可视化的曲面。
该项目的主要目标是研究各种重建算法,并从点云数据创建物体的三维模型。该项目从使用点云库创建三维多边形网格开始,理解其具体用法,并尝试采用各种算法以PCL(点云库)为基础从点云数据重建网格。PCL使用了许多算法,如行进立方体、网格投影、贪婪投影等,在对PCL的网格投影算法进行实验后,该项目重点是使用行进立方体算法创建网格。利用主成分分析和埃尔米特径向基函数,从点云出发建立了行进立方体数据。
Recent advancement in the three-dimensionalacquisition techniques provides huge input of unorganized three dimensionalpoint data sets. With an increase in the application areas making use of pointclouds, there is a growing demand to reconstruct a continuous surfacerepresentation that provides an authentic representation of the unorganizedpoint sets and render the surface for visualization. The main goal of the projectis the study of various reconstruction algorithms and the creation of a 3dmodel of an object from a point cloud. The project was started with the use ofPoint Cloud Library to create 3d polygonal meshes and understanding its usageand experimenting with the various algorithms used in PCL (Point cloud Library)for reconstructing meshes from the point clouds. PCL makes use of manyalgorithms like Marching Cubes, Grid Projection, Greedy Projection, etc. Afterexperimenting with the PCL’s Grid Projection algorithm, the project focused oncreating the mesh using Marching Cubes algorithm. The data for Marching Cubehas been created from the Point Cloud using Principal Component Analysis andHermite Radial Basis Function.
1 引言
2 问题描述
3 本文组织结构
4 重要术语
5 使用的工具与技术
6 研究背景回顾
7 贪婪投影算法
8 行进立方体算法
9 Kdtree:最近邻近似
10 使用主成分分析进行表面法向估计
11 埃尔米特径向基函数
12 投影结构与算法流程
13 结果与比较
14 面临的问题
15 结论
16 参考文献
下载英文原文地址:
http://page2.dfpan.com/fs/blcjb22112918681bc8/
更多精彩文章请关注微信号: