I'm a geologist and It's my first time to use Photoscan for aerial photogrammetry.
I've an exarotor drone and I've tried to scan a disused quarry. My problem is that the edges of quarry are not on the same level but there ise a difference of 30 meters.
Usually the flight is done while maintaining the same level, but in this case if you were flying, for example, 30 meters from the edge higher while maintaining the same proportion would observe the lower limit of 60 meters losing much detail.
I want to change flight altitude as you step from the higher elevations to the lower ones so you do not lose detail.
Are you trying to maintain the same resolution throughout your entire area? In that case, that depends more on your camera and your UAV flight altitude than it does for Photoscan. Photoscan in general applies more of a computer vision structure from motion approach to solve the problem of making a 3D model from photos, as compared to a more photogrammetric approach. The difference is that Photoscan will inherently be able to take into account these differences in height as it is trying to figure out where the camera should be to make information on automatic tiepoint matching 'fit' For the Canon ELPH520 HS cameras that we are using, the difference in estimated ground sampling difference or pixel resolution between 30m and 60m above the ground is about double - 1.26 cm vs 2.5 cm. If you need to maintain the 1.26 cm resolution throughout your mission, you would want to have the drone changing altitude, but you will also need to account for photographic overlap, as this is significantly reduced with changing altitude and a very important constraint for Photoscan reconstruction. I have found that counter to traditional photogrammetry texts, the structure from motion photogrammetry approach, Photoscan included, tends to need > 75% forward and side overlap.