To find the poses of the tags on the body, you would typically just specify one of the tag's poses with respect to the body.
Then you have
- a few tags for your fixed world frame ("lab") where you specify at least one pose of which
- a camera rig (that's another body) which is non-stationary
- the body (which is no stationary).
Now you turn the body around such that tagslam can discover all tags on it.
When you do the "dump" service call, the output file will have the missing tag poses of the tags on the body.