Ask Your Question

ArUco, Position estimation, Blender

asked 2018-10-11 11:27:06 -0500

technoprog gravatar image

hi community,

i've been struggling for a week with the problem that i can't get the ArUco position data into Blender. I can't figure out my mistake. The calibration worked well because the values for tvec look very good.

My problem: I photographed the Aruco marker in the following position image description

the tvec values i receive are as follows: [0.07369, 0.18815, 24.88399] these sound plausible. For rvec I get [0.00833, 2.82668, 1.37063] (in degrees 0.47708°, 161.95675° 78.53137°)

Now the problems begin, I try to place a 3D model with the found alignment at this position with blender. But the found angles don't match with the ones I expected.

In blender I can only specify a rotation around x / y / z.

My tactic was to set the camera in blender fix to point 0 and move the 3D model accordingly. For the position it fits. If I rotate it around the found angles, the model does not stand as it does on the picture (which is also clear with these angles).

Now I ask myself, what I am doing wrong or what I have to do with the rotation vector.

I am currently writing my bachelor thesis and am therefore really grateful for any help. I'm already getting the crisis although I'm still doing the preparation work for my thesis.

edit retag flag offensive close merge delete

1 answer

Sort by » oldest newest most voted

answered 2018-10-11 11:50:35 -0500

dpizzle gravatar image

updated 2018-10-11 13:30:11 -0500

It looks like you are expecting rvec to be euler angles in radians. You need to convert rvec to a matrix first using cv::Rodrigues. And then convert the matrix to euler angles. Searching this forum for rvec and Rodrigues results in examples.

edit flag offensive delete link more


ah ok if I understand you correctly, the three angles obtained are not at all radian angles. (or do they? but what do they represent then (yes an angle but with witch connection to the object)?). I have already found the Rodrigues formula, the tip with the change to Euler angle is new for me and have I read it now and try it out. Tanks. I think the tip just helped me find another mistake. Up to now I have always rotated around the world axes (at the right point but rigid in the direction), but if I understood that correctly it must continue to rotate around the already rotated object axes. Thank you very much

technoprog gravatar imagetechnoprog ( 2018-10-11 14:28:42 -0500 )edit

As I understand it, the rvec is a pointing vector, and the length of the vector (in radians) corresponds to the radians of rotation in the plane normal to that pointing vector.

opalmirror gravatar imageopalmirror ( 2018-10-15 13:21:50 -0500 )edit

Question Tools



Asked: 2018-10-11 11:27:06 -0500

Seen: 300 times

Last updated: Oct 11 '18