AI surveillance has been a new favourite for crowd management in India, particularly at urban events and pilgrimage sites, to enhance safety and efficiency. However, its success has been inconsistent.
AI-powered systems are capable of analysing crowd density by identifying congested areas and surges, potentially preventing stampede-like situations. These systems use facial recognition to identify missing people and also detect any unusual or suspicious activities through behavioural anomaly detection.
There have also been instances of governments deploying drones and aerial surveillance systems to monitor large areas, manage traffic and crowd movement.
Tarun Wig, co-founder of Innefu Labs, an AI-driven company offering advanced data analytics and information security solutions for national security agencies, explained their surveillance systems in detail. “Our system continuously learns from baseline crowd behaviour. A sudden spike in velocity, erratic motion vectors, or group-level directional shifts trigger real-time alerts for possible panic or stampede scenarios, differentiating them from regular high-footfall events, such as rallies or festivals.”
He also added that the company deploys convolutional neural networks (CNNs) and Transformer-based vision models for crowd density estimation. Behavioural anomalies are detected using unsupervised learning models, such as autoencoders, along with recurrent neural networks (RNNs) that track temporal deviations from typical movement patterns.
IoT drone systems and predictive analytics are being increasingly deployed for a wide range of events and locations, from temples and festivals to airports and major sporting events.
However, while AI in crowd surveillance has matured relatively quickly in India, offering certain benefits, there is an urgent need for better governance frameworks, data rights protection, and accuracy standards, as well as measures to protect people’s privacy.
Interestingly, Innefu’s surveillance stack is designed with open APIs, enabling seamless integration with third-party drones, thermal sensors, acoustic detectors, and other IoT components in smart city environments. “The data is harmonised via our central AI fusion platform for real-time decision-making.”
Innefu’s AI-driven crowd surveillance system is also built on a multi-modal image and video analytics engine that combines computer vision, behavioural AI, and video intelligence.
This system ingests real-time video feeds from CCTV and raw footage, applying AI models for facial recognition and gait analysis. For intent detection, it monitors micro-expressions and behavioural deviations to identify potential high-risk individuals before an incident occurs.
The most recent deployment of similar systems was seen during the Rath Yatra in Ahmedabad on June 27, where an estimated 1.5 million people were expected to attend the procession. This gathering made the old city area one of the most congested spaces.
Recent stampede incidents in India, including RCB’s victory celebration in Bengaluru that caused 11 deaths and injuries, have prompted urban cities to enhance security with AI.
For the first time, Ahmedabad’s security measures included AI-based surveillance systems. According to PTI, the AI-based system could notify the police and the fire department in the event of a blaze and alert the police control room to potential overcrowding along the route. As per the reports, around 14-15 lakh people attended the Rath Yatra.
“Under the AI system, the visual analytics software at the police control room would receive live feeds from CCTV cameras and drones deployed on the route. The software will analyse the live feed to estimate the number of persons present at that particular spot and how many more are expected to arrive in a short period,” as reported by the PTI.
On June 27, the pilgrim city of Puri also saw over 275 AI-enabled CCTV cameras and drones deployed alongside 10,000 security personnel, supported by a real-time WhatsApp chatbot for public assistance, and an integrated command and control centre for centralised coordination and data-driven decision-making to manage nearly 1.5 million devotees, the press release said.
An Odisha-based defence tech startup, IG Drones, partnered up with the state government for “advanced aerial surveillance and counter-drone systems to ensure seamless monitoring and threat neutralisation during the event.”
Ahead of the event, Bodhisattwa Sanghapriya, founder & CEO of IG Drones, said: “This year, by integrating our advanced drone surveillance with anti-drone defence systems, we’re helping ensure that the Yatra remains peaceful, secure, and uninterrupted. Our aim is to [assist] authorities with real-time situational awareness and actionable insights that uphold the safety [of the public].”
However, despite these measures, the government saw yet another fatal failure in managing the crowd. Three people died and over 50 were injured during the Rath Yatra in Puri on Sunday morning.
IG Drones did not immediately respond to AIM’s request for comments.
Similarly, the AI cameras that monitored the crowd during the Maha Kumbh in Uttar Pradesh, in January this year, were a failure. Reports stated that more than 30 people died, and 60 sustained injuries after a stampede broke out at the snana site.
But why are these AI systems failing? The challenges are aplenty: Many tier 2 and 3 cities lack the backend infrastructure (bandwidth, compute power, and trained staff) needed for effective real-time analytics; inconsistent CCTV quality hampers model performance; facial recognition often struggles with darker skin tones, crowded settings, and partial obstructions; besides false positives and negatives can lead to misidentification, especially in sensitive community contexts.
India also lacks a comprehensive law to protect personal data. AI systems often collect identifiable data without consent, raising concerns about potential misuse and the tracking of individuals outside crowded spaces.
The post Puri Stampede: AI Fails Crowd at Management, Yet Again appeared first on Analytics India Magazine.
