适用于 3D 建模管道的文件格式和工具
我正在研究点云数据的建模/重建算法。到目前为止，我一直在使用 Python 进行开发，并且对 VPython 满足我的可视化需求感到相对满意。
我遇到的一个问题是，当渲染大量对象时（至少在我的非 3d 加速 Linux 笔记本电脑上），VPython 变得相当慢，使得对复杂模型的目视检查变得相当困难。
我一直在尝试使用外部工具进行可视化，但问题是我有点迷失在可能的文件格式和可用工具的海洋中。例如，我一直在尝试 MeshLab，它非常适合以简单的 ascii 格式显示点云数据，但我无法决定以哪种兼容格式导出其他类型的几何图形，以叠加在点云层上。
以下是我的整个流程的要求：
- 点云数据可能包含数百万个点，存储为简单的 xyz ascii 坐标
- 建模基元主要是直线和圆柱体（即没有多边形），数量以数千计
- 可视化工具理想情况下应该是交叉的-platform（它必须至少在Linux上运行）
- 应该有一个Python模块，可以轻松导入/导出所选文件格式的数据（或者格式足够简单，可以编写一个简单的转换器，如果没有的话）
我一直在谷歌搜索很多关于这个所以我对所有这些问题都有初步的答案，但在我的背景下，没有一个是 100% 令人满意的。任何帮助或建议将不胜感激..提前非常感谢！
I'm working on a modeling/reconstruction algorithm for point cloud data. So far I've been developing in Python, and been relatively happy with VPython for my visualization needs.
One problem I have is that VPython becomes quite slow when rendering a great many objects (at least on my non-3d accelerated Linux laptop), making visual inspection of complicated models quite difficult.
I've been trying to use an external tool for visualization, but the problem is that I'm a bit lost in the sea of possible file formats and available tools. I've been trying MeshLab for instance, which works great for displaying point cloud data in simple ascii formats, but I couldn't decide in which compatible format to export my other types of geometry, to superimpose on the point cloud layer.
Here are the requirements for my whole pipeline:
- The point cloud data may contain millions of points, stored as simple xyz ascii coords
- The modeling primitives are primarily lines and cylinders (i.e. no polygons), numbered in the thousands
- The visualization tool should ideally be cross-platform (it must run at least on Linux)
- There should be a Python module for easy data import/export of the chosen file format (or the format is simple enough to write a simple converter, if not)
I've been googling a lot about this so I have tentative answers for all of these, but none that is 100% satisfying in my context. Any help or advice would be greatly appreciated.. many thanks in advance!
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（2）
我最终选择了 Geomview：查看器本身足够强大，并且它实现了许多 OOGL 文件格式满足我的需求。我对点云数据使用
.off格式，对其他建模基元使用.skel格式。这些文件格式也是人类可读的，这使得编写导入/导出函数变得容易。I finally settled for Geomview: the viewer itself is powerful enough, and the many OOGL file formats that it implements answer my needs. I use the
.offformat for point cloud data, and.skelfor my other modeling primitives. These file formats are also human-readable, which makes writing import/export functions easy.Panda3D 怎么样？它是跨平台的，只要你有一个像样的显卡，它应该能够处理数百万点的渲染。
How about Panda3D? It's cross-platform, and it should be able to handle rendering millions of points as long as you have a decent graphics card.