Video are not best source for CC you won't get location or orientation from video without using surveyed or fake controlpoints. Video doesn't have any camera metadata at all. The best would have the already precalibrated P4 camera or at least take one picture before taking video. You could get away without control points but then will have to input at least for 3-6 images orientation and position coordinates. So workflow is as follows, import video it will extract frames from each second as pictures then go trough all the pictures and delete all blurry photos as it will cause issues with aerotriangulation. Also delete all frames where UAV is rotating(because of rolling shutter artifacts). If the afterwards emaining photo count is small then you will have to set to extract frame each 0.5sec.
↧