0

After running RQDecomp3x3 in OpenCV, you get:

mtxR – Output 3x3 upper-triangular matrix.
mtxQ – Output 3x3 orthogonal matrix.
Qx – Optional output 3x3 rotation matrix around x-axis.
Qy – Optional output 3x3 rotation matrix around y-axis.
Qz – Optional output 3x3 rotation matrix around z-axis.

How do you get back from the three rotation matrices (Qx, Qy, Qz) to the original input matrix?

Or in the case where the input matrix was a rotational matrix, mtxR will be the identity matrix so how can you go from the three rotation matrices to mtxQ?

UPDATED With answer though I don't get why the transpose is needed.

Alex Rothberg
  • 10,243
  • 13
  • 60
  • 120
  • transpose of a rotation matrix is same as inverse – Micka May 10 '17 at 20:25
  • original matrix should be mtxQ * mtxR, but I know it as QR decomposition, not RQ decomposition?!? So your question shoulf be how to get from Qx, Qy, Qz to mtxQ? – Micka May 10 '17 at 20:52

1 Answers1

0

It looks like (at least for a rotational matrix input): input = (Qx @ Qy @ Qz)'.

Alex Rothberg
  • 10,243
  • 13
  • 60
  • 120