CloudCompare forum

Hello everyone,

I am a recent user on this software. Very pleased with it. I use it to exploit data coming from a Zeiss Comet machine (on parts for the automotive industry). A colleague asked an interesting question though : is Cloud Compare in accordance with the ISO norms that governs dimensional measurements. I'm thinking for example about the algorithm that does the best-fit registration.

Any information on this subject would be welcome. The final answer for me would be to know if I can use this software professionally, in accordance with the IATF (https://www.en-standard.eu/iatf-16949-2 ... -industry/)

Does anyone use this software in the car industry, aeronautics... ?

I don't know ;). And I don't have access to this standard (as it's not free). Do you maybe have the excerpt about plane fitting?

We use the standard 'ACP' method (Eigen values f the cross covariance matrix of the points).

Hello Daniel,

Thanks for your answer. I'm french but I guess we'll stay in english for the others.
I have sadly very few info to give you. I'm a technician in a metrology lab, I for example use a CMM to perform dimensional measurements and we of course have to use software that performs computation according to the associated norm (gaussian, circumscribed, inscribed and Chebyshev association criteria for example).
I use now Cloud Compare to also perform dimensional measurements, so we were wondering about the ISO norm part as well...

CC would not conform in all places. Using it in industry should be tested explicitly for the use case you want and accredited for a subset of the full program. I have tested a number of the functions against a NIST SRD for point cloud shapes, I can't find the link to the SRD I used (it has been a while since I downloaded it sorry).
CC has a number of edge cases in cloud to mesh measurements that would generally preclude it from most "certified" use cases.

Many of the functions are rock solid and will perform well against standard test data sets.

Any of the cloud to primitive distance methods have worked for me.
The sphere center/radius fit worked compared to expected results.
The plane fit worked compared to expected results.

I would say your quality org will need to define a test procedure for your use case to validate locally. plus this is open source so you would need to keep your validation limited to specific versions you test as changes can come commit-to-commit

Thank you, that is very useful.
Didn't know about NIST, even less about NIST SRD.

In our case, sets of comparison with the results given by polyworks (https://www.innovmetric.com/en/products ... nspectortm) could be useful.