Project in collaboration with Hitachi Rail STS. Teams of students coming from different backgrounds and fields of engineering. The main aim of the project is to create a system of analysis and management of pedestrians in delimited spaces, with reference to critical scenarios and control decisions to be made in long term and real time when the flows are considerable. The role I assumed is of responsabile of all the code used for the project. In particular, My team proposed an IOT system able to collect data from surveillance cameras by using an object detector learning built with MobileNet+SSD technology. The system is able to do an optimisation of the pedestrian flow inside a building by using some parameters obtained with surveillance camera. Furthermore, it is able to inform people, inside the building and outside, about the level of congestion, in real-time.