CloudCompare forum

Not with CloudCompare... You'd better use a simpler tool for that (even though I don't know which one :D).

I guess you try to use CSF? (if yes, make sure to use the latest 2.13 version as the algorithm has been improved). And then, regarding the Noise cleaning or SOR filters, it's generally a trial and error process to set the parameters correctly. But in the end it's always a trade-off between how much ...

By training a classifier dedicated to it? (as it's very linear and peculiar). Of course it depends on how dense the point clouds are.

Don't hesitate to share a sample cloud with me if you want me to try something (admin@cloudcompare.org).

So technically, the Poisson Reconstruction algorithm is already a command line tool (as provided by its authors: https://www.cs.jhu.edu/~misha/Code/Pois ... sion16.04/).

It's as is the command line mode was not loading the I/O plugins... You should see the list of plugins loaded at the beginning of the logs (in both the GUI and the command line versions).

Ah ok. It looks a lot like the 'countour plot' option of the Rasterize tool.

So I would try to remove the non-ground points (with CSF, scissors, etc.) directly from the raw cloud. And then use the 'Rasterize' tool.

Good catch. This should be fixed in the next 2.13 minor release.

Hum... you are correct! You'll have to set a high value for now.

I'll fix that for the next minor release of the 2.13 version (should be out in a few weeks).

And was the global shift sufficient? (just in case...)

Ah, maybe that's the LAZ format that's not supported? Does it work without the end part? (-EXT LAZ)