Revision history [back]

Several input parameters to the OpenCV calls you are making will need to be modified so that they apply the the reoriented (rotated 90 degrees) input image: cameraMatrix (K), calibCoeffs (D), and imgSize must be modified to apply to the new image orientation.

See https://docs.opencv.org/master/dc/dbb/tutorial_py_calibration.html for a starting point of resources to look into.

• K is simple, exchange fx with fy, and cx with cy.
• D is more complicated - the p1, p2, and p3 coefficients are affected by rotation (I don't have a ready answer on those).
• imgSize must of course be changed to match the new number of rows and columns.