Example: 60 sec video recorded with 30 fps. Option ”Video desired FPS” = ”1”.
Extracts the first frame of every second, resulting in 60 images to process instead of 1800. 1 frame per second is often a pretty good overlap and you reduce so much processing while still being able to use video instead of photos. IMO video is more simple when mapping in manual flight.
I’ve seen that you can use softwares for this but having it implemented so you can choose that 1 video would be awesome. I dont know how hard it would be to implement but i’ve seen that Roboflow has this feature when you import video to create a dataset. It’s open source so maybe someone with expertise could use that as help?