In order to set the transformation between the different frames in our robot system, we must publish the tf message of such transforms. The most appropriate and generic way to do so consists of using the camera_pose stack; we will use the latest version from this repository, which can be found at https://github.com/jbohren-forks/camera_pose. This stack offers a series of launch files that calibrate the camera poses with respect to each other. It comes with launch files for two, three, or more cameras. In our case, we only have two cameras (stereo), so we proceed this way. First, we extend camera_stereo.launch with the calibrate_pose argument that calls calibration_tf_publisher.launch from camera_pose:
<include file="$(find ...
