是否有任何预先编写的库/软件来存储和可视化点云？
我目前正在开发一个动作捕捉 3D 扫描系统，并且正在生成一堆点云，每个点云的大小介于 9megs/文件之间（存储大约 600k 点，每个点有 4 个整数，xyz 强度） ）。
摄像机以每秒 60 帧的速度进行捕捉，目前，我们希望捕捉 3 分钟左右的视频，因此每个视频文件最终会接近 100 GB。
我们对这些数据有两个目标。不太重要的是，我们想看看是否可以采取任何措施来无损压缩数据。但更重要的是，我们希望能够以相当不错的帧速率在计算机上可视化这些数据（希望具有所有旋转和缩放功能，但只需设定相机角度就可以了）。
与此相关的主要问题是将大型数据集加载到内存中以及如何管理所有这些、创建简单的 GUI 以及将点本身可视化。
我想我可以写一些在这方面做得不错的东西，但我真的不想重新发明（方形）轮，而且由于我对此没有任何经验，我会向你们询问一些建议（语言、库、框架等）。
提前致谢！
I'm currently working on a motion capture 3d-scanning system, and I am generating a bunch of point clouds, each of which are somewhere between 9megs/file (storing around 600k points, with 4 integers, x-y-z-intensity, for each point).
The cameras are capturing at 60 frames per second and currently, we are looking to capture something like 3 minutes of video, so that'll end up being something close to 100 gigs per video file.
There are two objectives we have with this data. Less importantly, we would like to see if there is anything we can do to compress the data losslessly. More importantly though, we would like to be able to visualize this data on the computer at a somewhat decent frame-rate (hopefully with all the rotating and zooming goodness, but it would okay with just a set camera angle).
The main concerns with this is loading a large data set into memory and how to manage all this, creating a simple GUI, and visualizing the points themselves.
I think I could write something that does a decent job at this, but I don't really want to reinvent the (square) wheel, and since I don't have any experience with this, I would ask you guys for some suggestions (languages, libraries, frameworks etc).
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（1）
嗯... VTK 用于可视化；你会得到你想要的所有交互性，而且 600K 点在一个像样的 GPU 上听起来并不算太多（即使是 4 年历史的 NVidia 7 系列或同等 ATI 也应该可以处理）。
至于无损压缩，我建议对数据使用二进制格式，然后将其压缩（例如，使用 zlib 或其他东西）。
Hmm... VTK for visualization; you'll get all the interactivity you want, and 600K points doesn't sound like too much on a decent GPU (even a 4-year old NVidia 7 series or equivalent ATI should handle it).
As for lossless compression, I'd suggest using a binary format for your data and then just zipping it up (say, using zlib or something).