why invert pose

asked 2020-07-12 13:15:56 -0500

superfly gravatar image

I have a pose output from solvepnp(), not a prob. Many sources suggest to take the inverse of the pose. I am having a contextual blindspot here: can someone tell me why I should use the inverted pose. In common sense language how is the inverted coordinate space more accurate/more appropriate.

edit retag flag offensive close merge delete


Read this or this about homogeneous transformation

Eduardo gravatar imageEduardo ( 2020-07-13 07:07:55 -0500 )edit

Thank you, read this before and successfully modeled it. What I m getting at is less theoretical: what will inverting pose will do; how does the world look with an inverted pose versus normal pose -- any examples?

superfly gravatar imagesuperfly ( 2020-07-13 13:54:06 -0500 )edit

Many sources suggest to take the inverse of the pose.

No idea why.

solvePnP estimates the pose of the object in the camera frame. The inverse homogeneous transformation gives the pose of the camera in the object frame.

It is simple like that.

Eduardo gravatar imageEduardo ( 2020-07-13 22:57:51 -0500 )edit

This may be a nomenclature issue for me. When you state "camera frame" do you mean camera space coordinates? And when you state "object frame" are these coordinates sourcing from the object?

superfly gravatar imagesuperfly ( 2020-07-14 15:16:24 -0500 )edit