Closed 7675t closed 5 years ago
Thanks, very thoughtful to not affect current behavior.
@7675t it seems that the marker's orientation is not changed, only the tf changed
@zhudelong Not sure what you mean to say. I believe #65 is on the difference between the marker and tf orientation, and this PR's argument is to make them identical.
@7675t sorry for not explaining it clearly. I use the parameter you provided
<param name="rotate_marker_axis" value="false"/>
Then use the following commands:
rosrun tf tf_echo d435_color_optical_frame camera_marker
rtopic echo /aruco_tracker/markerer
they give the consistent result about the orientation and position, but when I use RVIZ to show the TF and the marker. I find the marker plane is still aligned with x-0-z plane.
@7675t
by the way, do you know how to get the corner coordinates directly via the topics? Thanks a lot!
@zhudelong Still I'm not quite sure... what did you use to display the marker (red square plane) on rviz? I don't think aruco marker has its own rviz plugin. TF looks correct, so it depends on the marker visualization.
And about the corner point position, there is no topic to publish them. You have to drag out them from aruco's marker detector if you need.
BTW, this is an closed pull request so you should open new issue or ask your question in ROS answers. Thanks.
@zhudelong Sorry my fault, you mean you are using simple_single node and it publish an visualization marker.
And its y scale is 0.001, that's why it looks wrongly oriented.
Sorry, I will fix it.
@7675t Got it. Thanks a lot! I should have noticed that.
This add a boolean parameter named
rotate_marker_axis
for #51 . Default is true and it does not effect current behavior. When it is set false, the unnecessary rotation is skipped and the orientation of tf should be identical to the marker detection results.