Preliminary Checks
Proposal
Hi there, was looking for the following feature, have not been able to find any docs related, if any, please share them.
Add launch parameters to set the frame_id used in the image and pointcloud message headers, so they can be published directly into the calibrated frame without any post-processing.
Use-Case
In robot deployments, the camera's physical pose rarely matches the TF frame published by the driver exactly. After running a hand-eye or extrinsic calibration, the corrected frame exists in the TF tree, but since the driver hardcodes its frame_id, all image and pointcloud data must be republished into the calibrated frame, adding unnecessary overhead.
Allowing users to set the frame_id via a launch parameter (example from other camera driver OrbbecSDK_ROS2 #58) would let the data land in the right frame from the start, which would be appreciated.
Anything else?
No response
Preliminary Checks
Proposal
Hi there, was looking for the following feature, have not been able to find any docs related, if any, please share them.
Add launch parameters to set the
frame_idused in the image and pointcloud message headers, so they can be published directly into the calibrated frame without any post-processing.Use-Case
In robot deployments, the camera's physical pose rarely matches the TF frame published by the driver exactly. After running a hand-eye or extrinsic calibration, the corrected frame exists in the TF tree, but since the driver hardcodes its
frame_id, all image and pointcloud data must be republished into the calibrated frame, adding unnecessary overhead.Allowing users to set the
frame_idvia a launch parameter (example from other camera driver OrbbecSDK_ROS2 #58) would let the data land in the right frame from the start, which would be appreciated.Anything else?
No response