How to Compare Multiple Point Clouds

[BrokenArm54]

Hello,

I am a new user of CloudCompare.

Coming soon, I have to compare point clouds made by scanner over a longer or shorter period of a machine in a factory to know the degree of subsidence of the whole because of the vibrations. This sag should be highlighted by comparing scatter plots and I should edit a report.

Could someone tell me what would be the best procedure to perform the day I should scan the engine?
How to compare point clouds in CloudCompare?
How to highlight the differences between point clouds?
How to measure these differences?
How accurate would these measures be?
Is it possible to edit a report of all these calculations?

Thank you in advance for your help.

[daniel]

Ah ah so many questions at once ;)

And most of them have answers either here on this forum, or on the wiki (https://www.cloudcompare.org/doc/wiki/i ... =Main_Page).

Basically everything is possible, but the final report (but you can create labels on specific points, and of course export either some particular points or the whole set of distances for analyzing them in an external tool).

[BrokenArm54]

Hello,

Oh yes ! Many many many questions about this amasing tool lol !

Thank you for your reply.

[BrokenArm54]

Hello,

I made a first simple test at my office using the blk 360 scanner. As with the attached image snapshots, I first scanned a construction cone paused on a table and then scanned a second time the same construction cone but this one moved a few centimeters.

then, I imported the two point clouds into CC in pts format. With the naked eye, I can see that the cone of building site moved but my goal is to highlight its displacement by highlighting the area of ​​displacement and to know how much it has moved. So I used the tool> distances> cloud / cloud dist ...
But at this point, although having tested the different color scales, I can not isolate the area of ​​displacement. Is it possible to highlight it by isolating it? Because it would allow me, in my opinion, then to know more easily how much the cone of construction has moved. I hope to be pretty clear in my thinking. The idea of ​​a test, in the end, is to be able to use the same procedure to highlight disorders of a few millimeters on structures.

Thank you.

[BrokenArm54]

2 Point Clouds

[daniel]

You can disable the 'show NaN/out of range values in grey' checkbox in the 'Parameters' tab of the scalar field (in the cloud properties section). Then by playing with the 'min displayed' and 'max displayed' interactors of the active scalar field, you can isolate the points with large distances.

Another option (less interactive, but more accurate) is to use the 'Edit > Scalar fields > Filter by value' tool.

(Mind also that the C2C distance computation tool is simple. For more accurate/robust measurements you could use the M3C2 plugin).

[BrokenArm54]

Hello Daniel,

I will test the M3C2 plugins and I will come back to you for any questions. Thanks again for all this valuable information.

Have a nice day.

[BrokenArm54]

Daniel,

J'ai regardé le plugin M3C2 et j'ai sous échantilloné le nuage de base comme conseillé puis j'ai coché Horizontal pour les normals. Je vois effectivement plus facilement les points déplacés du cone. Par contre est ce normal que la table et les chaises apparaissent alors qu'elles n'ont pas été déplacées? Est ce peut etre due un a probléme d'alignement des deux nuages de points?


I looked at the M3C2 plugin and I sub-sampled the base cloud as advised then I checked Horizontal for normals. I can easily see the displaced points of the cone. By cons is it normal that the table and chairs appear when they have not been moved? Could it be due to a problem of alignment of the two points clouds?

[BrokenArm54]

Piece jointe