连接点以形成三角形
如果我有一组 3d 点(又名点云),那么确定我应该创建的 3 点(三角形)组以创建表面重建的最佳方法是什么?
If I have a set of 3d points (AKA point cloud) what is the best way to determine the groups of 3 points (triangles) I should make, to create a surface reconstruction?
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(3)
Delaunay 三角测量 是您的朋友!如果你用谷歌搜索这个术语,就会有很多关于它的资源,而且它背后的数学/逻辑并不是太难。使其快速有点困难(但完全可行),但这完全取决于您的要求。
Delaunay Triangulation is your friend! There are lots of resources available about it if you Google the term, and the math/logic behind it isn't too tough. Making it FAST is a bit harder (but totally do-able), but that depends entirely on your requirements.
如果您愿意使用外部应用程序(或实现自己的应用程序),MeshLab 有很多很酷的重建算法下:过滤器 >重新划分网格、简化和重构。 球枢轴表面重建和泊松重建是不错的选择。
If you are willing to use an external application (or implement your own), MeshLab has lots of cool reconstruction algorithms under: Filters > Re-meshing, simplification and reconstruction. Ball Pivoting Surface Reconstruction and Poisson Reconstruction are good ones.
Vierjahn 等人提出了一种新方法。 sGNG:基于生长神经气体的在线表面重建。尽管实现起来是一个相当复杂的算法,但它允许对 3D 网格进行增量细化。
您可以在此处阅读技术报告。此外,作者还给出了提示并参考了 3D 点云 3D 网格重建研究领域的其他新技术。
A new approach has been introduced by Vierjahn et al. sGNG: Online Surface Reconstruction based on Growing Neural Gas. Although it is a rather complex algorithm to implement it allows incremental refinements of the 3D mesh.
You can read the tech report here. Furthermore, the authors give hints and refer to other new techniques in the research field of 3D mesh reconstruction from 3D point clouds.