Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

First thing to be clear, there is no perfect calibration board detector, so that even with the right settings in OpenCV etc. there is no guarantee that OpenCV will always detect the calibration board although it is present in the image.

That aside, you should verify that your board settings are correct, make sure you have the right number of rows and columns etc. Besides I noticed there is a shading gradient across the board, as if it was being illuminated with a single point light source. This is not a good idea with the OpenCV chessboard detector. I have noticed it is sensitive to the relative difference between white and black tiles, if the difference is too low, i.e. because of poor illumination, it will not detect the board. I suggest you take images of your calibration pattern in a room with plenty of ambient illumination.

Last but not least, in the image it is apparent that your calibration pattern is not flatly attached to the board. This should really be fixed. The camera calibration routines assume that the board is perfectly planar, when this constraint is violated your calibration results will be arbitrary, they may be good, but they may also be poor.