Danish Airs and Grounds: A Dataset for Aerial-to-Street-Level Place Recognition and Localization
This provides a new benchmark for researchers working on aerial-to-street-level localization, though it is incremental as it builds on existing datasets with more diversity and size.
The authors tackled the challenge of place recognition and visual localization under wide baseline conditions by introducing the Danish Airs and Grounds dataset, which includes over 50 km of road imagery with accurate 6-DoF metadata, and they proposed a map-to-image re-localization pipeline that uses dense 3D reconstructions from aerial images.
Place recognition and visual localization are particularly challenging in wide baseline configurations. In this paper, we contribute with the \emph{Danish Airs and Grounds} (DAG) dataset, a large collection of street-level and aerial images targeting such cases. Its main challenge lies in the extreme viewing-angle difference between query and reference images with consequent changes in illumination and perspective. The dataset is larger and more diverse than current publicly available data, including more than 50 km of road in urban, suburban and rural areas. All images are associated with accurate 6-DoF metadata that allows the benchmarking of visual localization methods. We also propose a map-to-image re-localization pipeline, that first estimates a dense 3D reconstruction from the aerial images and then matches query street-level images to street-level renderings of the 3D model. The dataset can be downloaded at: https://frederikwarburg.github.io/DAG