You Are Here: Geolocation by Embedding Maps and Images
This improves geolocation for applications like navigation or mapping by generalizing a previous semantic feature approach with higher accuracy and faster convergence.
The paper tackles geolocalizing panoramic images on 2D maps by learning an embedded space to compare images and map neighborhoods, achieving over 90% accuracy for routes of about 200m using Google Street View and OpenStreetMap data.
We present a novel approach to geolocalising panoramic images on a 2-D cartographic map based on learning a low dimensional embedded space, which allows a comparison between an image captured at a location and local neighbourhoods of the map. The representation is not sufficiently discriminatory to allow localisation from a single image, but when concatenated along a route, localisation converges quickly, with over 90% accuracy being achieved for routes of around 200m in length when using Google Street View and Open Street Map data. The method generalises a previous fixed semantic feature based approach and achieves significantly higher localisation accuracy and faster convergence.