Unexpected output from EstimatePoseSingleMarkers()

When detecting a generated (flat) aruco marker, if I convert the rvec to a matrix with Rodrigues, should this be a identity matrix (or close to it)?

Is there any other way to ‘debug’ rvec or the rotation matrix?

“flat”? facing the camera?

not quite. the coefficients should be close to 0, +1, -1, but permuted, because you’re having a rotation of 180 degrees (marker faces you) and whatever rotation around the optical axis.

Is there any other way to ‘debug’ rvec or the rotation matrix?

What I usually do is to draw the different coordinate frames, for example the rotation matrix from marker coordinate frame to camera coordinate frame:

image

1 Like