AI-Powered People Counting for Social Distancing and Transport Safety

Demos Architecture
IREX Crowd Video Analytics estimates the number of people in a camera field of view or any zone of interest. The counting accuracy depends on how well the camera can see people. The average accuracy is similar to that of a human brain counting people on a still frame.

Real-Time Alerts
The IREX platform sends real-time alerts upon the actual number of people reaching a predefined threshold and stays above it for a predefined period of time. The real-time alerts are used to detect:
  • Violations of social distancing policies
  • Dangerous crowding on public transport (e.g. on railway platforms, escalators)
  • Oversized queues  (e.g. et entrance gates, cashier tills, ticket offices)

IREX can export people counting data to third-party data analysis software, for example, Google Data Studio. The statistics data can be used to analyze passenger flows and optimize staff working schedules, public transport schedules, and traffic lights.

Two types of foot traffic data are available:
  • The total number of people from the Crowd Video Analytics
  • The number of people crossing a virtual tripwire in either direction from IREX rule-based video analytics

The following demo shows 3 parameters to generate real-time alerts for the venue manager and security staff:
  • Number of people: 6 persons
  • Minimum detection time:120 seconds 
  • Detection interval: 10 seconds