Edited and so it works, with the exception of now losing all of the CBlobResults includes when linking. I'm going to figure that out next.
Added a way to spoof target position.
It's an rviz interactive marker which can be controlled by the mouse,and which will publish 3d target positions. There are still some TODO'sin src/rviz_interactive_target.cpp.
The code compiled! Now to make sure everything does what it's supposed to do.
Nick made some changes to make it compile.
Created the vision package
For now there is a skeleton of v2v3_converter (without the math) whichwill take in a TargetDescriptor (x, y, size) and tf from /camera to/kinect, and will output a 3d target location.