Wrong pose of aruco tag after upgrade to ros noetic

[Hmueller-hako]

We have migrated our software from ROS Melodic to ROS Noetic. I noticed that the pose from the Aruco tag is no longer correct. Also, the coordinate system in the result image is drawn in the wrong position (similar to #89 (comment)) .

With version 2.0.2:

With commit 52c8c2b :

I use the following launch-file:
`

<arg name="markerSize"      default="0.11"/>    <!-- in m -->
<arg name="side"             default="front"/>
<arg name="ref_frame"       default="camera_$(arg side)_depth_optical_frame"/>  <!-- leave empty and the pose will be published wrt param parent_name -->


<node pkg="aruco_ros" type="marker_publisher" name="aruco_marker_publisher_$(arg side)_camera">
    <remap from="/camera_info" to="/camera_$(arg side)/color/camera_info" />
    <remap from="/image" to="/camera_$(arg side)/color/image_raw" />
    <remap from="~/markers" to="/aruco/markers" />
    <param name="image_is_rectified" value="False"/>
    <param name="marker_size"        value="$(arg markerSize)"/>
    <param name="reference_frame"    value="$(arg ref_frame)"/>   <!-- frame in which the marker pose will be refered -->
    <param name="camera_frame"       value="camera_$(arg side)_depth_optical_frame"/>
</node>

``` The camera_info-message `rostopic echo -n 1 /camera_front/color/camera_info header: seq: 13528 stamp: secs: 4155 nsecs: 142000000 frame_id: "camera_front_depth_optical_frame" height: 480 width: 640 distortion_model: "plumb_bob" D: [0.0, 0.0, 0.0, 0.0, 0.0] K: [319.9988245765257, 0.0, 320.5, 0.0, 319.9988245765257, 240.5, 0.0, 0.0, 1.0] R: [1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0] P: [319.9988245765257, 0.0, 320.5, -0.0, 0.0, 319.9988245765257, 240.5, 0.0, 0.0, 0.0, 1.0, 0.0] binning_x: 0 binning_y: 0 roi: x_offset: 0 y_offset: 0 height: 0 width: 0 do_rectify: False --- ```

[saikishor]

Hello @hacmue ,

Thank you for showing interest in the package. Could you please test the same by passing a rectified image and by setting the param image_is_rectified to True, let us know if it fixes it?, why is your camera frame pointing to the camera_depth_optical_frame instead of the camera_color_optical_frame?.

It seems using the rectified image this doesn't happen:

However, When using a non rectified image in the simulation, we could reproduce the issue you are mentioning

We would try to fix it in the upcoming weeks after some internal planning, however, if you can open a PR, we can take a look.

Thank you,

Best Regards,
PAL Robotics Team

[saikishor]

Hello @hacmue,

We have found the bug and fixed it. Probably, tomorrow or next week you will have fix here in GitHub. Thank you for reporting the issue.

Best Regards,
PAL Robotics Team.

[saikishor]

Hello @hacmue,

We have just updated our repository with the fix. We will be closing this issue as the fix has been deployed. We thank you again for pointing out the issue. In case of queries, feel free to open the issue again.

Thank you,

Best Regards,
PAL Robotics Team.

[hacmue]

Thanks a lot. It works fine.