Auto-match data points

[stevef]

We use the "Auto-Match data points" command some times to force a set of measured hole centers to match the nominal hole centers.  I noticed that this command takes a very long time to process if there are more than a few dozen points being matched.  Any comments on this?  Is there a work around to speed up the process?

[prehistory]

Yeah, Auto-Match is really designed for a few dozen points or less.  The IMInspect Reference Guide says to use it for no more than 200 combined points, but I've used it successfully with more than that.  Two things really speed it up.  One is to make the Tolerance setting as low as possible, based on your best guess of nominal to measured deviation.  The second is to lower the number of points in the Initial Match Optimization (found under the Advanced section of the interface).  In my experience, setting it low (e.g. 10) doesn't change the ultimate fitting result, though if some of your points are way off, it may cause the alignment to fail completely.

Anyone else have thoughts on this?

[tobiwan]

we used auto-match-data points to align a faro arm in a field of about 100 reference points by probing a few of these reference points.
This might be a bit off-topic but shows a nice features of the function: the amount of data points in the source object and the target object can differ significantly.