Machine for Rhino Pointcloud Processing

Dear Rhino whisperers,

I’m struggling with some processes in Rhino 6 (also trying WIP) and am wondering if I can speed it all up by investing in a better machine.

I’m trying to open point cloud data in PLY format, then remove unnecessary points to a separate layer (or deleting). I then contour/section and loft through the rebuilt series of curves. Once I have an accurate nurbs surface I export it and everything is fine. The editing of the pointcloud is extremely painful.

2

These PLY files are approximately 200-500mb (the last is 458mb) and when resaved as 3dm they become much bigger, like 2.2gb.

This is my v7 WIP SystemInfo systemInfo.txt (2.0 KB)
and v6 systemInfoV6.txt (2.5 KB)

In an ideal world, I’d have a machine that could help process pointclouds faster and Vray to render a bit faster too. At the moment a 1000/1000 pixel render is taking the best part of 12 hours. Penguin is also very slow.

this is my current machine:

machine

Any advice would be much appreciated,

thanks,

Giles

Hi Giles - I guess I’d invest in more RAM, at least 32 GB if you are messing with point clouds of this size. Massive point clouds are handled better in general as we go forward, but these are not Rhinos strong points. I have a script someplace that will divvy up a point cloud into boxes as divisions of the overall bounding box and shove each cloud on its own layer - that might be helpful - only showing a portion at a time to work on.
Shout if you think that might be useful - I can tell you it will not be fast on a huge cloud- time it with a significant coffee break.

Is this in Vray? I’d check with their forums - settings make all the difference.

-Pascal

Hi Pascal,

Yes it’s Vray’s trial period. I suspect I have too many/too complex lighting. But that wasn’t the case with Penguin. It just seems to take ages. I will trawl their forum.

The script you mention could be really handy. So yes please!

I just tried CloudCompare last night and it helped a lot. The file is now manageable within Rhino.

thanks,

Giles

Here you go - see if that does anything good.

PointBoxer.py (4.6 KB)

To use the Python script use RunPythonScript, or a macro:

_-RunPythonScript "Full path to py file inside double-quotes"

-Pascal

CloudCompare is nice for pre-processing point clouds before bringing into Rhino.

https://www.danielgm.net/cc/

thanks both. I’ll try the script Pascal. Perhaps a combo of CC and your script will work magic.