How Bengaluru's AI-powered system reduced late-night firecracker violations this Diwali: Report

Bengaluru police successfully employed an AI-enabled monitoring system to address late-night firecracker usage during Diwali, significantly reducing complaints.

In a fresh development, Bengaluru police said they were able to curb late-night firecracker use during Deepavali this year, crediting an AI-enabled monitoring system for resolving roughly 41 per cent of complaints lodged after the 10 pm cutoff.

Bengaluru police utilized an AI monitoring system to reduce late-night firecracker use during Diwali, addressing 41% of complaints post 10 pm. (Representative image) (Pexels)

According to Kuldeep Kumar Jain, Joint Commissioner of Police (Administration), cameras positioned at key points around the city were trained to identify firecracker explosions and smoke in public areas, said a report by The Hindu. Once detected, alerts were immediately forwarded to the central command centre, which then dispatched Hoysala patrol teams to the spot. Jain noted that this rapid-response setup played a major role in cutting down noise levels and improving air quality during the festival.

This marks Bengaluru’s first deployment of artificial intelligence specifically to curb night-time firecracker violations, the report said.

The AI-driven Firecracker Detection and Response Initiative has clearly shown that the Bengaluru Safe City Project can incorporate advanced AI solutions for real-time civic enforcement, Jain said, as quoted by the publication. He added that the collaboration between the city police and tech firms resulted in better on-ground awareness, more accurate enforcement, and enhanced public safety.

The operation made use of the city’s extensive surveillance ecosystem, which includes over 7,500 cameras linked to the Integrated Command and Control Centre (ICCC). More than 200 of these cameras, located in densely populated neighbourhoods such as Srirampuram, KR Market, HSR Layout, Haralur, and Marathahalli, were analysed using the Firecracker Detection Video AI app from Awiros Appstack, powered by the Awiros Video Intelligence OS, the report noted.

The AI was calibrated to spot tell-tale signs of firecracker use, such as bursts of light, plumes of smoke, and unusual spikes in crowd activity. When such activity surfaced on camera, the system automatically created an alert packet containing the precise location, time, and relevant video snapshots. These notifications were sent instantly to the ICCC and to ground units through WhatsApp, allowing operators to verify the report and coordinate a swift police response.