We will need to run robotic simualtions that feed the vision system in order to test the vision recognition in multiple "real-world" configurations. 

When can we off-line program an application such as the one in the video? 

EXAMPLE VIDEO

Comments?