Towards a Real-Time Big GeoData Geolocation System Based on Visual and Textual Features

摘要

During the recent years, new sensors and new methods of collecting geospatial information emerged. This technological advancement is behind the apparition of Big GeoData concept. The exploitation of such concept demonstrates its effectiveness in various activity sectors. However, it poses new challenges that cannot be overcomed using traditional solutions. We believe that a good exploitation of Big GeoData can be guaranteed by establishing, first of all, a good placing strategy yet called geolocation estimation activity. In fact, generating efficient geospatial knowledge is closely linked to positional accuracy. Since that traditional geolocation estimation methods are not able to handle Big GeoData variety, a new solution has to be proposed. This is the object of the present article where the architecture of a real-time multimode approach is described.