There was an ESRI (ArcGIS) conference in San Diego a couple weeks ago - videos of the plenary sessions have just come available.
Notable is the UAV section of the mid-morning session - fast forward to minute 78:
The connection is possible because of a metadata stream referred to as "MISB" (http://www.gwg.nga.mil/misb/faq.html). I wonder how difficult this stream of data is to glean or convert using APM2 telemetry and a fixed camera? (I suspect camera position is also worked into MISB data)
Has anyone tried this? I'd like to try with the municipality I work for - but would love to not have to reinvent that wheel, also...