Computer vision to estimate amount of people in public spaces

indeed similar and Sarah focuses more on the outcome (use case) whereas I was focusing more on the solution. we could be a good team. Webcams would be a great source but it will depend on the space (usecase) we want to monitor. Yes, we need deep learning but the technology went super far and it's quite easy to use with some technical skills. Forecasting is certainly possible but we need more historical information to catch seasonality. The ability to know in advance the number of people at a given space in a given time would in fact be very interesting for the general public. Our project could focus on this as the final outcome. To recap: 1. use cameras and potentially webcams (most precise way of monitoring); 2. keep it private, no privacy concerns (assure the audience we don't do facial detection nor sell data to authorities and 3rd parties); 3. define a concrete usecase (location, audience); 4. forecast (predict occupancy)

Hello Luka, thank you for your idea. It is very similar to the idea of Nathalie and Sarah. https://participate.smartcity.brussels/fr-BE/ideas/taux-d-occupation-des-espaces https://participate.smartcity.brussels/fr-BE/ideas/redirection-vers-une-autre-zone-si-l-espace-est-trop-frequente-via-une-app But Sarah wasn't found to use camera, expensive and not friendly. What do you think? Do we need deep learning? What about a simple image of a webcam, that gives people an indication: empty / some people / nice crowd / too crowded? And last, how could we predict the amount of people?