# solvepnp(), transformation output from which coordinate system?

The documentation of solvepnp is a bit unclear to me. Is the output tvec (translation) the distance from the checkerboard to the camera coordinate system? And rvec the rotation of the camera coordinate system relative to the checkerboard coordinate system?

In other words how can I use the rvec and tvec to get the position of the checkerboard frame in the camera coordinate system?