Accurate geolocation (luna-location? seleno-location?) remains a hard problem on the Moon, and makes Terrain Relative Navigation a non-trivial task to this day.
I present here Galatea, a fused Visual Place Recognition model that can make existing models perform closer to terrestrial models levels of accuracy on extra-terrestrial surfaces.
Accurate geolocation (luna-location? seleno-location?) remains a hard problem on the Moon, and makes Terrain Relative Navigation a non-trivial task to this day.
I present here Galatea, a fused Visual Place Recognition model that can make existing models perform closer to terrestrial models levels of accuracy on extra-terrestrial surfaces.