How to assign an image_message to a transformation
Hi dear TF community.
I am publishing two image messages, "depth_image" and "infared_image" from two cameras. My goal is to register the depth image on the ir img. I have made the external calibration, and do publish it using the static_transform_publisher. The internal calibration parameters are published from the camera nodes. I am visualizing the images in rviz, and do see the two frames I created (depth_camera and ir_camera). Now if I switch in rviz between the frames, nothing happens to the images. My guess is, that I haven't assigned the frames to the cameras.
So... If my guess is right, how would I assign a frame to a camera node? Or if not, do you maybe have an idea what is missing for the registration process?
Thanks, Josh.
Hi @max_ros take a look at prior question: https://answers.ros.org/question/2449...
Yes you will need to assign transform for each camera
Hi @osilva. Thanks. I followed the hints in the answer you linked and am now on the following status:
However I still haven't archieved my goal, which is to overlay the ir_image on the rgb image (or on the pointcloud, doesn't matter). Do you have a hint for me how to archive this? Do i need to apply the tranfsormations to the images manually? (I though rviz takes care of that for me?).
Hi @max_ros check if this is what you are trying to do: https://answers.ros.org/question/3217...
Settings: https://github.com/robogeekcanada/STE...
Hi @osilva - yes this was again a very good hint, thanks!
However I still haven't archieved my goal :-D I now use the depth_image_proc register node, which is supposed to map my depth map on my infrared image. From there on I could also project both on a pointcloud. This in general works fine, except that I only do receive black images. So I strongly guess something is wrong with my calibration. However I do have troubles to debug this. Do you by chance also have a hint how to proceed (is is the internal calibration of the infrared camera, is it the external calibration, ...)?
Hi @max_ros I suggest we wrap this question with a partial answer based on the progress you made so far. Document your effort so others may benefit and start a new question with the new things that are still unresolved. This way more people will have visibility to help as well. Also with a new question, you could more pictures to help visualize what you are dealing with.
@osilva : new question