M3C2 - Normals orientation

[Teresa]

Hi,
I'm a surveying student and working on my master thesis. I'm doing a deformation analysis of an historic monument. The data was aquired with TLS inside the monument, once a year for 5 years.
I am currently using the M3C2 algorithm to compare my point clouds.
I'd like to know, which parameter is better for the orientation of the normals: +/-X, +/-Y, +/-Z, +/-Barycenter, +/- (0,0,0).
I have already compute the data with different settings and the results are quite similar but some spaces are different.
Has somebody an experience with M3C2 with a point cloud taken from the inside of a building?
Thanks !

[daniel]

Well, it all depends on the building shape of course. But I would say by default '+Barycenter' if the building has a convex shape.

Probably the best option is to compute the normals before using M3C2 (with Edit > Normals > Compute) and check that the result is fine. See https://www.cloudcompare.org/doc/wiki/i ... %5CCompute

[Teresa]

Thank you for the quick answer !
Indeed, +Barycenter seems to be the best setting for my data.