I have a point cloud created from a B&W Leica laser scan at 27 locations within a factory and around 1000 images taken with a Nikon D3000 DSLR (the photos have no position data). I'm having trouble aligning both sets of data. I'm hoping to produce an accurate and visually impressive model by combining both sets of data. Is there a way to do this without any sort of physical reference in both measurements?Thanks,Cameron
I do this by picking coordinates of "structure" within the point clouds and applying those coordinates as GCP's in the photos. I assume you have a "registered" point cloud - so all the clouds are on the same coordinate system .. you'll have an E57 ready to roll in to CC.
A factory will have heaps of "structure" to pick from .. centres of bolts .. intersections of corners in 3-planes of surfaces .. corners of ducts etc ..
I tend to find I have photogroups of "runs" of images - 50 to 100 at a time perhaps - I tend to find it a challenge to get a great AT .. also check the "hints" in photos you haven't nominated control in, that have control - see how good the hints are compared to the GCP .. soon tell you if the AT is any good.
Thanks, this method seems to work quite well! Not quite perfect yet but I guess I could add some more GCP to improve this further.Many thanks