Thanks for the info. Currently, we are heavily dependent on auto registration. We are noticing that we tend to get a big difference between cloud data and survey data (secondary check) over a large number of scans. We are trying a different method where we geo-reference strategic scans, then do the automatic registration with all other scans when these are in place. Have you used this method? What another method would assist us to get better accuracies?
Are these projects small like a building or spread out like a road or city? Just tying to find out if the mismatch is because of scale factors or because of the registration just not matching up.
I have tried a mix a target based registration for a base and then tying the other scans to that base and it has worked well but for overall accuracy I prefer to use targets. However some argue that they still like the targetless. I would say if you go full targetless to get a large amount of overlap between scan stations.