Cloud to Mesh Artifacts

[Linzoid3000]

Hi Daniel,

I've been doing C2M, making a mesh of a high quality SfM derived dataset.
All of the results (several clouds compared to this mesh) have what appear to be errors of geometric shapes.

I used the same reference data to run DoD and M3C2, and the artifacts are not occuring in the results.
This leads me to believe that the mesh is the source (the las cloud doesn't have them, nor the DEM).
I tried to have a closer look in CC, but zooming in on the mesh crashes the program.
I reproduced the mesh in ArcGIS using las to TIN, and the artifacts aren't there either.

My last thought is that it could be a manifestation of the C2M itself.
I have read that in the mesh's topography is heterogeneous, the distance can be confused when the angle normal to the mesh is not generally towards the other cloud...if I understand correctly, this is the advance of the M3C2, as it doesn't use the angle normal to a continuous mesh surface....

I've attached a screen grab. The correct data is red, and the geometric errors are blue. The range between these colours is almost 4meters....
any thoughts?
Thanks!
Lindsay

edit: The attachment quota has been met? 13KB?

[daniel]

Generally the limit for atached images is more on the resolution (I believe it's 1000 pixels max for the largest dimension). Otherwise you can send me the image by email (cloudcompare [at] danielgm.net). You can also send me the data if it's not too big.