It's actually not too difficult. The camera turret has incredibly precise GPS and knows exactly what angle it's at. With that, it's pretty straightforward to figure out where it's pointing at on the ground, grab the relevant map data, and overlay graphics.
I wouldn't call it "straightforward" but it is definitely computable. I worked on such a cartography-video overlay system about 10+ years ago, and you need at a minimum: precise GPS (as you pointed out), distance offset between GPS antenna and camera gimbal, precise aircraft attitude in all three axes, precise gimbal attitude, and a very good terrain model giving you altitudes of the land around the camera target. You also need to project your location, camera vector, and field of view polygon into a coordinate system that takes into account the curvature of the earth, because these aircraft are often many km from their targets. Very interesting project.
