yaw roll pitch from extrinsic parameter

Not much experience in the general case, but curious where the camera
is and what its looking at.  More detail might give better answers.

For example, when using overhead cameras, I've found a few brightly
colored markers on each wheeled robot to be adequate.  I'd imagine a
flying camera could make fair progress with a simple horizon detector.
  If you're looking at a plane, the traditional red/white/green
navigation lights might be helpful.

Once you've found a set of keypoints in the image, its then just
linear algebra and optimization to find coordinate transforms which
match...

- Daniel

thanks for your suggestion,
it is very simple and general, assuming your camer looking down at
ground plane...
and assuming, you've got the extrinsic camera matrix...the Rotation +
translation coefficients...
Hence, from 3x3 matrix, can we know these parameters : yaw, roll pitch
angles?

so not to worry about what pattern to look at...
thanks