Snapshot navigation (SN) algorithms use stored egocentric views to drive visual navigation in a manner that is thought mimic the excellent navigation of small brained insects. Most reported SN algorithms use panoramic images and have been shown to work well when close to a memorised route or place, but are less capable of navigation when the agent is displaced from the training locations. Here we introduce an alternative with two visual Fields of View (FoV) analogous to storing a snapshot from the left eye when facing rightwards of the goal direction and vice versa. Using a virtual environment, based on a natural desert ant habitat, we created a grid of insect realistic visual inputs. For two navigation tasks (place navigation and route navigation) we calculated the direction of highest familiarity towards using memories pointing towards the central place and along the route, for both the single panoramic view (Cyclops) and for the binocular methods (Braitenberg). Virtual agents using the two methods were deployed to follow the direction of highest familiarity of the closest grid point and their success at arriving at the goal location or completing the route were recorded. We found that using two FoV vastly increases the successful navigation for both taks, in comparison to using a single panoramic view.