Curvy object

[Adam]

Hi Daniel,

If you have a curvy object like the one below, how do you fit them? I normally use Fit>plane but it does seem working for this. I also have tried others but still not working. What is the best fit? I have looked at a curvy object where is the best fit is quadric, but I have tried it as well and it doesn't look right.

Thanks

[Adam]

And if there is a marker on the object does this affect the volume or M3C2 calculations? for example, if one of the objects is marked with a marker pen and the other not.

thank you

[daniel]

Well, I'm not sure what your problem actually is: the 'Fit plane' fits a plane (flat, planar) on the selected cloud. The fitting relies on the Least Squares formulation (i.e. minimization of square distances).

The Quadric fit, will try to fit a 2.5D Quadric (see the equation here: http://www.cloudcompare.org/doc/wiki/in ... it_Quadric).

I'm not sure what you were expecting.

And regarding your "marker", if it has a 3D shape, then yes it will certainly change the volume or distances calculation even in a small way. If it's only color, then no :| (pay attention that some scanners don't measure the same distance depending on the color, therefore it can result in a small bump).

[Adam]

The problem is I get a very curvy surface, so I'm not sure what fit to use for the surface I have uploaded.I was expecting the surface data to be corrected as sometimes the scanner not facing the surface well.

[daniel]

Well I guess the Quadric should do the job. Otherwise there's no other option in CC anyway.

[Adam]

Regarding the volume. Will the volume estimate the volume loss from the old scan to the new one? so it tells me how much has been lost or added? and how the distance is calculated? is it different from M3C2?
how about the std (standard deviation) from M3C2 plugin (it will estimate the roughness of each scan?
If there is a volume loss in the new scan, do I expect STD cloud 2 to be less than STD cloud 1 or this is a totally different method? because I don't most of the time. How can I relate these two, the volume and the STD cloud 1/2?

Thank you

[daniel]

Computing the volume change with M3C2 is highly approximate. Moreover the roughness is mainly used to determine whether the detected change (distance) is significant or not. Therefore I'm not sure it will influence the volume in the way you expect.

You should definitely use the 2.5D volume calculation tool here. And before that you just have to convert the meshes to point clouds (with 'Edit > Mesh > Sample points').

[Adam]

Thanks. what I meant is what is the realtionship between the volume (as the volume calculation has its own way to calculate the distance and will estimate what has been addes or removed) and STD cloud 1or 2 from M3C2 plugin (means the local roughness of each clouds which also has its own way to calculate the distance and estimate if cloud 1 had become rougher or smoother than cloud 2?). So if the volume has increased it doesnt mean that the STD cloud 1 should increase (higher) than STD cloud 2? please do correct me if I'm wrong.

Is the cross section tool another way of segmentation? I found it helpful to segement the edges in the right way (with measurments)

Regards

[daniel]

Sorry I don't really see any straightforward link between the roughness and the volume (there must be one, but it's not the only parameter that will change the volume estimation! (far from it in the general case).

[Adam]

Thanks. Is the cross section tool another way of segmentation? I found it helpful to segement the edges in the right way (with measurments)