From October 1, 2025, Delhi will see the rollout of 3,500 Artificial Intelligence (AI)-enabled surveillance cameras, advanced gunshot detection systems, and public alert mechanisms under the first phase of the Safe City Project. This initiative aims to enhance city-wide safety, with a strong emphasis on protecting women in public spaces and improving emergency responsiveness.
The decision to accelerate surveillance efforts follows a comprehensive audit of vulnerable areas across the National Capital Territory. The audit was commissioned in response to past safety lapses and revealed the need for approximately 10,000 additional cameras to ensure better visibility and quicker intervention capabilities. The first phase focuses on high-risk zones and densely populated areas identified during the study.
The project involves the deployment of various high-tech camera systems: 6,121 bullet cameras, 1,622 Automatic Number Plate Recognition (ANPR) cameras, 370 Facial Recognition System (FRS) units, and 1,876 Pan-Tilt-Zoom (PTZ) cameras. These will be supported by 200 gunshot detection units, each comprising three acoustic sensors, and 300 public address systems to support real-time communication with the public.
Unlike conventional surveillance systems, these AI-driven devices will use Artificial Intelligence Video Analytics (AIVA) to detect gunshots, monitor abnormal crowd gatherings, track stolen or hotlisted vehicles, and identify suspicious activities. Real-time alerts generated through these analytics will be processed at the Integrated Command, Control, Communication and Computer Centre (C4i) for swift verification and escalation.
Once validated, alerts will be relayed to Deputy Commissioner of Police (DCP) offices and local police stations for immediate action. This tiered system ensures that security personnel are equipped with timely, actionable information to respond efficiently to incidents.
A major feature of the Safe City initiative is its focus on women’s safety. The AI software is designed to detect distress patterns, such as a woman being isolated, encircled by a group, or showing signs of panic. These events trigger high-priority alerts, prompting Police Control Room (PCR) vans to respond proactively, even in the absence of a formal complaint.
Areas such as hostels, PG accommodations, college campuses, and workplaces with high female footfall are being prioritised. In addition, the system monitors crowd density, flagging gatherings of more than 50 people as high-priority and smaller groups as lower-risk, enabling the police to respond accordingly.
AI capabilities will also be used to monitor abandoned objects, activate PTZ cameras in response to gunshots, and match vehicle license plates with law enforcement databases. Each event is recorded with a timestamp and high-definition visuals, aiding evidence collection and court proceedings.
The system also enables post-incident audits, allowing senior officials to review footage, assess response times, and improve future training and accountability. Alerts determined to be false are discarded with documented reasons, while valid alerts are logged and responded to in real time.
This technology-driven approach is expected to revolutionise urban policing. The remaining phases of the project are expected to be implemented by the end of 2026, eventually covering the entire city under a unified AI-backed security framework.