This paper presents a geo-localization framework of street-level outdoor images using multiple sources of overhead reference imagery including LIDAR, Digital Elevation Maps and Multi-Spectral Land CoverUse imagery. We describe five different matchers and an adaptive linear fusion process which combines individual matchers probability maps into a single map. These matchers exploit mountain elevation profiles, rendered camera views, landmarks, landuse classes and building heights. We successfully validated our framework on 100 queries with geographic truth in two world regions each of 10, 000km squared in the USA.
2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , 23 Jun 2013, 28 Jun 2013,