CloudCompare forum

Hi!
I need some information about tool cloud2cloud distance.
You can see the project in this link:[www.dropbox.com/s/64rm7u2i37z843l/INFO% ... .docx?dl=0]https://www.dropbox.com/s/64rm7u2i37z84 ... .docx?dl=0[/www.dropbox.com/s/64rm7u2i37z843l/INFO% ... .docx?dl=0]

From the results, how are considered the negative points by CC? Are they internal or external from reference point cloud? Did CC has a sign convention? If my point cloud is a plan and not a surface, how CC considers internal or external?
I thank you in advance and sorry for all these questions!
Chiara

Hum, in this case you shouldn't have a signed distance (it should be unsigned). I believe it's a bug ;).

Do you also get signed distances with the other local models ('least square plane' or 'quadric')?

Thank you for your rapid answer! :)
By the way, I've just tried but I've the same problem. What it means?

It simply means that it's a more general bug with all local models ;)

I'll fix this quickly.

Sorry for the confusion: there's no bug ;)

I misread your image titles, then I checked the code (everything was ok) and then I read your document again :D

The signed distances along X, Y and Z are simply the 'displacement' between the point and its nearest equivalent on the reference cloud (or on the local model in your case). Therefore they are simply expressed along the respective axis.

However with so few points in your reference cloud it's hard to deduce anything. Especially for the signed displacements (and even with local modelling the result can't be very accurate). You can compute the 2.5D Delaunay triangulation of the whole mesh (See the 'Edit > Mesh' menu). You'll see how coarse/inaccurate it will be.

Sorry! I'm reading now your reply!
Thanks you!