MARVEL

In the Municipality of Trento, the project will use the cameras and the microphones which are already available in the city; eventually, new equipment will be tested too.

The adopted system will be a shared “Edge-to-Fog-to-Cloud'' one which plans to do the audio and video analysis with AI algorithms directly “on edge” (in other words the analysis will be done close to the cameras), without using the Cloud. Especially during the experimental phase, some data will be used to “train” and continuously improve the algorithms thanks to Machine Learning (ML) techniques.

The project involves 17 partners of 12 European countries and it is coordinated by the Foundation for Research and Technology Hellas (GR). In addition to the Municipality of Trento, the Fondazione Bruno Kessler is also part of the Consortium as a technologic partner. The project is therefore in line of continuity with other numerous projects, applications and activities related to this topic in particular that are already carried out jointly on the territory. The total budget available through this project for the Municipality of Trento is 210.000€.

Use case: Municipality of Trento

The Municipality of Trento will represent a use-case to understand what happens in some areas of the city ( parks, squares, underpasses, road intersections) where cameras and/or microphones are installed. From the analysis of images and, where available, audio, the goal is to automatically recognise the presence/absence of people and to discriminate potentially problematic situations (gatherings of people, fights, illegal activities such as selling drugs) or traffic problems (accidents, congestion…). All this with the main goal of analysing the trend of these phenomena over time through analytics systems in order to be able to send real-time alerts to the Local Police station in case of potential dangerous situations.