I'm testing a block of image frames through the ortho mapping wizard. Frame table, camera table, CRS, and DEM surface are all accurate. But I do not have access to surveyed GCP, my footprint coverage is only about 50% overlap, 30% side-lap, and there is quite a bit of water frames or partial water frames.
So when I run the block adjustment tool, I'm only getting between 2 to 5-ray tie points. Even if I filter and delete all 2-ray tie points, my ortho-mosaic has huge shifts, both geometrically and radiometrically. It looks way worse than the pre-block adjustment mosaic (the mosaic derived from just the raw EO/IO).
Does anyone have any advice or coverage requirements, tie point requirements, how to best filter and manage tie points? Best block adjustment, seamline, color balance, and orthomosaic settings to use for a dataset like this? How to best deal with water tiles? Any advice would be appreciated. Thanks.