Wrong visualisation of point cloud after calibration

[TongZhangTHU]

Hello,

After I calibrated the RealSense camera extrinsics relative to the Franka Emika Panda using easy_handeye, I visualized the point cloud in the robot frame with rviz /camera/depth/color/points topic (I have published the extrinsics TF), but found that the point cloud is in the wrong position and looks rotated in some way, I suspect that it is caused by the wrong coordinate system. I wonder if you have any suggestions for my problem?

[marcoesposito1988]

Hi @TongZhangTHU,

did you visualize the point cloud after starting the system including publish.launch, or is still using the dummy position of the camera from the calibrate.launch file?

[TongZhangTHU]

@marcoesposito1988

I visualize the point cloud from the /camera/depth/color/points topic after starting roslaunch realsense2_camera rs_camera.launch filters:=pointcloud.

[marcoesposito1988]

Yes, but after doing the calibration did you restart everything, publishing the hand-eye calibration with this launch file, or is the system still running with calibrate.launch instead?

After calibration the new transform is not applied automatically, you have to restart everything using a different launch file to use the result of the calibration.

[TongZhangTHU]

@marcoesposito1988

The results I presented above were obtained after restarting everything and publishing the hand-eye calibration with this launch file, just as you suggested.

[marcoesposito1988]

It seems that the point cloud is rotated 90 degrees along two axes, so probably a different frame is being used to publish the calibration than what was used during the calibration. Please double check what is passed as frame for the camera

[Skeptic9793]

I'm also experiencing this problem, which seems to be caused by the fact that realsense's optical coordinate system definition is different from the ROS coordinate system definition