Hi All,
I work with opencv for blob tracking using gopro cameras for their wide angle lens. I use opencv's undistort and calibration functions but I wondered what the variation between lenses was. It is quite tricky to calibrate the gopro as it doesn't have a screen. I find I get quite a lot of variation depending on how close I can get the chessboard image to the edge of the frame.
Does anyone have any interest in in pulling together good values that could be used a starting point to remove most of the fish eye without taking calibration shots?