Page 1 of 1

Using registration markers for local best-fit

Posted: Wed Aug 25, 2021 3:54 pm
by cubafornian
I'm trying to use cloud compare to do some quality analysis of parts with point clouds/meshes generated by a 2-axis, blue light scanner. These parts can have small markers, such as a recessed plus sign (or whatever shape the community thinks is best) on different surfaces of the part. I'd like to be able to quickly select these features on both the reference mesh and the point cloud, and perform an align using one pair at a time. Ideally I would be able to script this so that every time a pair of features is selected, I can calculate and export the distances between the other correlating feature pairs all over the part. I understand CloudCompare may not be the best software tool to do this, but I'd like to stick to an open-source option if possible before moving on to GOM Inspect or the like.

Re: Using registration markers for local best-fit

Posted: Thu Aug 26, 2021 6:04 pm
by daniel
Well, you could use the 'Align' tool, if your 'markers' are spherical or hemispherical (and with a known radius). This way you just have to click on the spheres in both entities (in the right order).

See https://www.cloudcompare.org/doc/wiki/i ... itle=Align

Re: Using registration markers for local best-fit

Posted: Fri Aug 27, 2021 8:50 am
by cubafornian
Thanks, Daniel! I didn’t know it was possible to use hemispheres. We’ll have to give that a try! Is it possible to use a hemisphere and two other points to make a pair? That would be handy.

Re: Using registration markers for local best-fit

Posted: Sat Aug 28, 2021 8:08 am
by daniel
"two other points"? Not sure to understand sorry :D. You can pair a hemisphere (center) with a single point in the other cloud. And you can mix hemispheres and points as well.

Re: Using registration markers for local best-fit

Posted: Mon Aug 30, 2021 6:42 am
by cubafornian
Yes! Mixing hemispheres and points. Sorry my question was unclear. Do the hemispheres have to equal 180 deg? Or can they be less than half of a sphere? What is the tolerance exactly?

Re: Using registration markers for local best-fit

Posted: Mon Aug 30, 2021 12:43 pm
by daniel
Well, that's a best fit process. So it depends on the noise, etc. I guess it would work with less than a hemisphere if there's not too much noise.