Suitable file format and tool for 3d modeling pipeline

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 poi开发者_运维问答nts, 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!

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 .off format for point cloud data, and .skel for my other modeling primitives. These file formats are also human-readable, which makes writing import/export functions easy.

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.