Possibility to train canupo classifiers?

[KajN]

Hi!

I'm very new to CloudCompare and point clouds in general, but I've looked at some examples and tutorials of Canupo and it seems very useful.
The Swedish geography and mapping authority, Lantmäteriet, provides access to some open geo data, among them point clouds. The point clouds however, are only classified into water, bridges, ground, high/low and unclassified, which means both buildings and greenery are unclassified. So, my question is this: Would it be possible to use the Canupo classifier to separate these? I'm really only looking to use the buildings, to create context models for architecture and city planning. Or is there a better way to do this?

Demodata can be accessed here: https://www.lantmateriet.se/en/maps-and ... -datasets/
For the point cloud, scroll down to Elevation datasets, and select the Demo dataset from the row Laser Data NH, column City area. The file is a laz, around 190 MB in size. The same file is available here from my drive: https://drive.google.com/drive/folders/ ... sp=sharing

Kind regards
Kaj

[daniel]

Thanks for the links.

CANUPO is very powerful for vegetation. For buildings, it might work if most of the buildings have the same characteristics and shapes / size. This might be a challenge.