Unmanned aerial surveillance is key for the protection of archaeological sites against looting activities. Automated surveillance is a valuable decision-making means that can become of great support to public authorities. The European project Safeguard Heritage In Endangered Looted Districts (SHIELD) aims to developing a fully autonomous surveillance system that involves a drone capable of taking off and surveying a scene of interest (i.e., detecting/tracking objects, monitoring suspicious actions), as well as landing and charging at its smart helipad. In this extended abstract we focus on SHIELD perception capabilities, in particular on the the object detection module that is based on the state-of-the-art CenterNet algorithm. Preliminary results on thermal infrared data, from the BIRDSAI dataset, show promising results that could serve as a build-up of future improvements
SHIELD: Safeguard Heritage In Endangered Looted Districts
Mohamed Lamine Mekhalfi
;Nicola Saljoughi;Davide Boscaini;Fabio Poiesi
2022-01-01
Abstract
Unmanned aerial surveillance is key for the protection of archaeological sites against looting activities. Automated surveillance is a valuable decision-making means that can become of great support to public authorities. The European project Safeguard Heritage In Endangered Looted Districts (SHIELD) aims to developing a fully autonomous surveillance system that involves a drone capable of taking off and surveying a scene of interest (i.e., detecting/tracking objects, monitoring suspicious actions), as well as landing and charging at its smart helipad. In this extended abstract we focus on SHIELD perception capabilities, in particular on the the object detection module that is based on the state-of-the-art CenterNet algorithm. Preliminary results on thermal infrared data, from the BIRDSAI dataset, show promising results that could serve as a build-up of future improvementsI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
