Maha Kumbh 2025: AI-powered safety at six Prayagraj railway stations

The context

In early 2025, hundreds of millions of pilgrims travelled to Prayagraj for Maha Kumbh, one of the world’s largest religious gatherings — most of them by rail. With the railway authorities, Awiros deployed facial recognition and video analytics at six stations: Prayagraj Junction, Prayag, Sangam, Phaphamau, Jhunsi and Rambag.

The challenge

• Unauthorised access: heavy footfall made it easier to slip into restricted areas.
• Security incidents: more people meant more lost persons, theft, vandalism and tampering with cameras.
• Manual monitoring gaps: operators couldn’t track movement across every feed in real time.
• Slow emergency response: without automated alerts, people in distress were found late.

The solution

Facial RecognitionMatched faces against a secure database to find suspects and missing persons in real time.

Camera Tampering DetectionFlagged blocked, moved or damaged cameras to keep coverage unbroken.

Attribute-Based SearchFound people by clothing colour, gender and other features — key for locating missing persons.

Human & Vehicle TrackingFollowed people and vehicles across station premises in real time.

Fallen Person DetectionSpotted people who collapsed, for faster medical help.

Combination SearchCombined several attributes to find people or vehicles fast.

Loitering DetectionFlagged people lingering in sensitive areas.

Colour-Based SearchFiltered footage by colour to find specific people or objects quickly.

Intrusion DetectionCaught unauthorised access at platform ends and control rooms.

See it on your own sitesBook a demo

How it was rolled out

• Close collaboration: Awiros worked with station officials and system integrators to calibrate the system to each station’s layout and workflows.
• Control room integration: alerts and dashboards plugged into the existing control rooms.
• Real-time alerts: intrusion, tampering and fallen-person events reached security staff instantly.
• Built to last: the system stays in use well beyond the event.

The impact

• 70% less unauthorised access.
• 40% better passenger flow.
• Incident response cut by over 70%, from 10 minutes to under 3.
• 50% more efficient threat detection.
• 500+ missing persons reunited.
• 30% drop in vandalism and petty crime.
• 95% detection accuracy, with fewer false alarms.

The takeaway

Maha Kumbh set a new benchmark for safety and crowd management at mass gatherings — and gave railway authorities a model they can reuse at future large-scale events.