A question that has plagued me for years: what to set the camera parameters for a project with frames taken from a video as well as normal photos (particularly a drone). Obviously the video is lower resolution than the photos taken, so the same parameters wouldn’t apply. Do you group the photos into one calibration and one lens distortion group (as approximate) and the frames from video into another (with the same approximate parameters)? Do you put the frames from video as “unknown” for both? Do you set the same distortion model (like brown 3 with tangential) which works for the images to the video frames?
Please sign in to leave a comment.