When events are in closed spaces, it is enough to control entrances. However, in public events in open spaces where the number of entrances is too great, the capacity can only be controlled from the air. The use of drones is proposed as a suitable measure, due to their low cost and the possibility of flying at a low altitude.
Research is being carried out to find a suitable model of drone to carry out this kind of task and to improve the models of neural networks that exist and train them to recognise people from a bird’s eye view. Beyond recognising and counting people, the aim is to test a new model that can detect groups of people based on images provided by the drone of the local police in Castelldefels, who are collaborating with the project. The research team will work to optimise an algorithm based on artificial intelligence techniques that can be executed in real-time on a mobile device (a tablet or telephone) connected to the drone. All of this will be achieved while respecting the protection of people’s data and privacy. The images that are captures will be anonymised by automatically putting the faces of the people out of focus.
The project is coordinated by the ICARUS research group, with the participation of Castelldefels City Council.
The project was one of four UPC projects selected for the call ‘Retreating to grow: the impact of pandemics in a world without visible borders’ (PANDÈMIES 2020) of the Agency for Management of University and Research Grants (AGAUR), for research projects whose aim is to analyse the impact of the COVID-19 epidemic specifically, and the concept of pandemic in general. The projects will be developed in a maximum of 18 months.