One of the nice features that many cad based cmm software have is handling point clouds. If I try to think about this function in those old time software then the one pops up in my head that can do it better than others is the GON package in UX Umess. But now, many cmm packages provide good capability to deal with point clouds. And I'm not talking about point clouds from scanners with millions of points, that's a totally different subject, I want to focus on how to process the point data collected from your touch probe (either analog or touch trigger). Because many people don't realize sometimes you can utilize the scanning capability with the point cloud manipulation to make your inspection really fast. So I'll demonstrate this to you using OpenDmis software.

 I got a request from a customer to measure one die cutter to the drawing (no cad), about 30 dimensions (all 2D), +/- .001" tolerance, nothing fancy. I think everyone can do it without any problem, starting from alignment and then program all the features for dimensioning. But because I don't have the cad file to program with, so I kind of feel lazy to teach all the features and then fix nominals, doing alignment, etc. So all I did is scan the profile, which took me only 3 minutes from start to end, because it's a straight forward 2D scan, no need for any alignment. Then I saved the scan data to an iges file, imported it to OpenDmis and this is how it looks:

Now I activate the cloud operator by right mouse click the point cloud in the feature database:

1, I collected points from bottom edge to construct a best-fit line for alignment:

2, I collected points from left edge to construct a BF line then I can create my alignment:

3, Now I start to collect points for the features I want to dimension:

4, So from there I can report the radius, position, distance of these 2 circles and also I can even create a good graphical form error report since I got more than 100 points for each circles:

So you can see once I got a set of point clouds imported to OpenDmis, I pretty much can do whatever I want with these data to create a full report. This method has big advantage over conventional programming for the following reasons:

1, Minimum programming works on cmm (you might still need a good DCC alignment if dealing with more complicated parts), which means freeing your cmm for other works;

2, You get a quiet time to sit down and do a comprehensive analysis on your data;

3, All data is saved so you can always come back to it  if someone challenges you or you want to do some best-fitting, adding extra dimensions, etc.;

 The example I'm showing you is a very simple one, you can actually do a lot of more complicated manipulation on point clouds in OpenDmis.