Closed rohbotics closed 7 years ago
Yes, I agree. What do you think about having a single Fiducial message containing both vertices and transforms? It would fix an inconsistency that there is one message for each set of vertices, but all the transforms for a frame are published together.
I think we could refactor the messages to be more consistent. But I think that we should keep them separate in order to allow making the detection and pose estimation into separate nodelets in the future.
When there are no markers, we should publish an empty fiducial_vertices