CC doesn't work with Canupo

[emapodo]

Hi, I'm a beginner with CloudCompare. I opened a point cloud and followed the tutorial to use automatic classification with Canupo, but once I trained the classifier by applying it to a point cloud of not too large dimensions, CloudCompare lags and stops working. I tried with two different PCs, the problem persists even on the most performing one. What could it depend on?
Thanks for any help.
Emanuele

[daniel]

Which version are you using?

Anyway, it's probably due to a too large scale, that makes the classification process veeeerrrrryyyyyy long.

Try with smaller scales first (mind that the scales should be expressed in the same units as your point coordinates).

[emapodo]

I tried with 2.13.2 Kharkiv (08/07/2024) and now I'm trying with the latest version 2.14.alpha (18/06/2025). But even when I start training the classifier, Cloud Compare starts lagging and doesn't work. Attached are screenshots of the global shift/scale and trainer setup. Am I doing something wrong? I am trying to train the classifier to automatically classify the slab from the rest of the elements such as floor, walls etc.

[daniel]

I've tried on my side (with a random test dataset that I used in the past) and it still works properly with the latest 2.14.alpha version.

Things to pay attention to:
- you have a lot of scales (about 20) and the largest ones span over what looks like a third or half of the cloud width (which means it will encompass a lot of points, and therefore take a lot of time to compute)
- you may want to lower the max thread count (to 6 or 7) so that your PC keeps responding during the computation. I guess what happens is that the computation takes a very long time, and what you perceive as 'stop working' is that the processor is too busy.

If you think it's something else, don't hesitate to share your clouds with me (admin@cloudcompare.org)