Call me an idiot but I spent two days trying to figure out what should be obvious, but to me it's not: deleting tie points through selection>delete or gradual selection are still used when building a new mesh or point cloud. I understand why: depth maps are used, not tie points. But:
- Is deleting tie points only needed to reoptimize the camera? (should I reoptimize the camera at all since I calibrated my mavic 2 pro with the chessboard?)
- But then how to remove garbage baked into the mesh before it is created?
Masking photos as suggested
here is not an option: (a) too many small creeks (b) I want the textures and (c) no holes in my model. It seemed so simple: reflections from trees in the water appear as though underground so tie points can be easily deleted.
Looking forward to any clues, thanks.