There are numerous factors that might be potentially useful for large-scale localisation, including image retrieval, aerial to ground view image matching and building classification. Most of these aspects have not been thouroughly explored in the context of localisation. To that end, in SILDa, we focus on evaluating several such visual localisation related tasks.

The Scape-Imperial Localisation Dataset (SILDa) contains images across several conditions, and can be used to evaluate several tasks related to visual localisation.
The Scape-Imperial Localisation Dataset (SILDa) contains images across several conditions, and can be used to evaluate several tasks related to visual localisation.

More details, including the data and the evaluation protocols soon, please visit this website later.