I'm working on a 20K pictures project, token with a 5DS Canon in RAW and converted in JPEG 50MPX (5792 x 8688). Possibility to work on two different PCs
-pc #1 i7 3770K 3.50 GHz 8 cores - 32GB RAM - Nvidia Quadro K2000
-pc #2 i7 9750H 2.95 GHz 6 cores - 32GB RAM - Nvidia GeForce GTX 1660 Ti
I have decided to split into several groups.
The first group has 2700 pictures and it took 6/7 hours to get the alignment done on pc#1.
But I have a few problems in there
1) It did only about 1800 pictures on one component and several more components with 48, 13, or fewer pictures, as you can see above.
2) In the 1800 pictures one component, it puts point cloud in the middle of nowhere or part of them, bigger and moved from where they should be
3) Is it possible that with a 50mpx pictures RC doesn't recognize anything at that distance? I understand it would not be precise, but a minimum at least.
4) In the pictures I took, I have pictured even the floor from a different angle, why RC doesn't recognize them at all at some point? I have big holes, probably because it doesn't recognize enough points?
I still don't have aerial images due to bad wheater conditions to fly, can they solve the floor issue, problem #3 and a better overlap of all images?
I converted the RAW images even in 16 bit Tiff to have more information and after 10hours is still running with the same above settings (according to RC it would take other 24 hours)
What do I do wrong? There is another way to change something and see a faster result without wait so much time? Is draft alignment a good solution?
Please sign in to leave a comment.