Limited Period Offer:Be a PRO for 1 month @Rs49/-Multiple payment options available. Know More
you are here: HomeNewsPolitics

How a Noida-based exam security solutions firm is using AI tools to manage crowds at the Kumbh

The AI-based police surveillance system consists of 310 CCTV cameras, 50 internet protocol-based public address systems and five variable message display boards, apart from the computers, all of which keep an eye on the three lakh devotees who, on an average, throng the ghats every day

April 15, 2021 / 12:24 PM IST
A huge religious gathering in northern India could fuel a surge in Covid-19 cases, the government warned on March 21 as it called for an increase in testing and other health protocols. (Image: AFP)

A huge religious gathering in northern India could fuel a surge in Covid-19 cases, the government warned on March 21 as it called for an increase in testing and other health protocols. (Image: AFP)

Several million devotees are thronging the ghats in Uttarakhand's Haridwar to take dips in the Ganges during the ongoing Maha Kumbh congregation amid the deadly second COVID-19 wave sweeping the country.

While officials have admitted that they are struggling to impose safety protocols at such a big gathering, the Uttarakhand police have hired a Noida-based ‘examination security solutions’ firm to manage crowds as videos and pictures of pilgrims, squeezed shoulder-to-shoulder on the river banks jostling for a dip, emerge in utter disregard of social distancing norms.

Moneycontrol spoke with officials from the Uttarakhand Police and the Noida-based company to know how the Artificial Intelligence (AI)-based security tools have been used at the Kumbh.

The Company

The AI-based surveillance system, approved last year, was earlier supposed to be handled by the Hewlett Packard Enterprise, on the lines of an earlier project, the Dehradun Smart City Limited (DSCL).


COVID-19 Vaccine

Frequently Asked Questions

View more
How does a vaccine work?

A vaccine works by mimicking a natural infection. A vaccine not only induces immune response to protect people from any future COVID-19 infection, but also helps quickly build herd immunity to put an end to the pandemic. Herd immunity occurs when a sufficient percentage of a population becomes immune to a disease, making the spread of disease from person to person unlikely. The good news is that SARS-CoV-2 virus has been fairly stable, which increases the viability of a vaccine.

How many types of vaccines are there?

There are broadly four types of vaccine — one, a vaccine based on the whole virus (this could be either inactivated, or an attenuated [weakened] virus vaccine); two, a non-replicating viral vector vaccine that uses a benign virus as vector that carries the antigen of SARS-CoV; three, nucleic-acid vaccines that have genetic material like DNA and RNA of antigens like spike protein given to a person, helping human cells decode genetic material and produce the vaccine; and four, protein subunit vaccine wherein the recombinant proteins of SARS-COV-2 along with an adjuvant (booster) is given as a vaccine.

What does it take to develop a vaccine of this kind?

Vaccine development is a long, complex process. Unlike drugs that are given to people with a diseased, vaccines are given to healthy people and also vulnerable sections such as children, pregnant women and the elderly. So rigorous tests are compulsory. History says that the fastest time it took to develop a vaccine is five years, but it usually takes double or sometimes triple that time.

View more

Estimated to be a Rs 23 crore project, the contract was eventually awarded in December 2020 to Innovatiview India Pvt Ltd – a firm that deals with ‘redefining examination security’ to curb malpractices during examinations – for Rs 13 crore (including GST).

The firm will work with Uttarakhand police for a three-year annual maintenance contract (AMC).

Established in 2010, the company, according to its website, strives to support equitable and transparent conduct of exams and has a presence in more than 400 cities of India

“We are glad to be one of the first companies to bring an AI-based, high-speed, and extremely accurate touchless biometric verification solution to India,” announces the website of the firm, which is based in Sector 125, Noida, Uttar Pradesh.

READ: Kumbh Mela 2021: Here's everything you need to know before visiting

With a Rs 100 crore turnover, Innovatiview offers more than six security solutions and has so far served 75 clients with over 650 projects. Its list of customers—preceding the Uttarakhand Police—are Delhi Police, Chandigarh Police, the Central Board of Secondary Education (CBSE), and Staff Selection Committee (SSC)-Delhi, among other examination conducting agencies.

Said Avdhesh Sharma, project manager at Kumbh, Innovatiview: “We deal with all security-related equipment. Our major projects include Central Warehousing Corporation (CWC) and Union Public Service Commission (UPSC)/SSC examination centres, among others. We also handle the CCTVs installed in Delhi on January 26 and August 15. Recently, we were also involved in surveillance in red zone areas during COVID-19 in Delhi.”

CWC is a central government-owned warehouse storage and handling service provider.

The Technology

The system that functions out of a control room set up at the mela venue, consists of 310 CCTV cameras, 50 internet protocol-based public address systems and five variable message display (VMD) boards, apart from the computers.

Of the 310 cameras, 278 that are fixed are equipped with ‘varifocal’ lenses with zoom in and zoom out options. This enables law enforcement agencies to  captures the image of a person without a mask and send out an alert. Another 22 cameras are PTZ (pan tilt zoom) that can be moved in different directions while the remaining 10 are enabled with automatic number plate recognition (ANPR) system for vehicles.

The system is connected to I2V- an integrated video security software with video management system (VMS), built-in video analytics and options to integrate third-party devices.

Facial recognition works on context. Broadly, facial recognition is deployed for three uses. One, to establish an identity. Example: phones, that use facial recognition to unlock the home screen. The camera scans your face, turns it into biometric data and compares it to information it has stored. The same technology is also used by airports for automated checks. Two, to spot people in crowds. Here, facial recognition is used to scan groups of people and zero in on individuals. Three, to track behaviour. Some classrooms in China use this to spot unusual behaviour in children; for example, if they are in distress.

The technology that is being put to use at the Kumbh is a method of  verifying the identity of individuals using their faces captured real-time in the cameras.

“The whole analytics focuses on four major goals – ‘head count’, ‘object count’, ‘vehicle count’ and `identify people with no masks,” said Innovatiview’s Sharma.

‘Head count’ is to keep a tab on the number of people allowed to gather at a particular place to avoid overcrowding at ghats vulnerable to stampede. ‘Object count’ is to keep a tab on the actual number of people entering and exiting a ghat for the bath, at a time. ‘Vehicle count’ is used to calculate the number of vehicles that are moving towards the ghats or are in the parking area using ANPR cameras. The cameras are also meant to spot unattended objects. The public address systems can reach out to a 70-km radius, said officials.

“We followed the old, time-tested police methods for crowd and traffic management coupled with the latest AI technology. So, in a way it’s a blend of the two,” Sanjay Gunjyal, Inspector General (IG) mela, told Moneycontrol.

The Impact

The Kumbh Mela started on April 1 and Uttarakhand Chief Minister Tirath Singh Rawat inaugurated the surveillance control room on April 9.

The mela sees a daily average of over three lakh devotees at various ghats, but the number swells up during the shahi snans (major baths) held on March 11, April 12, April 14 and the last one, which is scheduled on April 27. On April 12, the total number of pilgrims who took bath was 31 lakh and includes those who also showed up on April 11 as well. Till 6 pm on April 14, as many as 13,51,631 pilgrims had taken a dip in Ganges at Haridwar marking  the Baisakhi snan.


In 2010, about 16 million pilgrims had taken this holy dip, known as Baisakhi snan, officials said.  Devotees believe that bathing in the Ganga will cleanse them of their sins and bring salvation.

Officials said that the AI technology was used primarily to avoid stampede-like situations across 107 ghats in the middle of the pandemic. Ashok Kumar, director general of police (DGP), Uttarakhand, told MoneyControl, that AI analytics did not come into play when things were running smoothly.

“We used this to zoom on faces and vehicles and identify people who are not wearing masks,” he said.

In pics | Rallies, religious gatherings aggravate India's worst COVID-19 surge

On April 12, Kumar had told India Today that it was not possible to impose COVID-19 protocols in such huge crowds.

Mukesh Thakur, assistant superintendent of police, Kumbh Mela, told Moneycontrol that the police used AI-enabled cameras to calculate that 21,34,000 pilgrims took bath on April 11, which came down to 9,98,000 on April 12.

“Over all 31 lakh pilgrims were present at the ghats on these two days. We also noted 68,000 vehicles on April 12 at the venue,” he said

As for identifying people without masks, Thakur said police avoided reaching out to pilgrims as it could have led to a stampede. Cops, however, distributed 1.1 lakh masks during the baths on April 11 and 12, he said.


Globally, the facial recognition system has its own up and downsides. A recently published BuzzFeed News investigation, ‘Surveillance Nation’, found that employees at law enforcement agencies across the United States ran thousands of Clearview AI facial recognition searches — often without the knowledge of the public or even their own departments.

Noted the Buzzfeed report: “For years, law enforcement agencies have experimented with facial recognition, a technology that promises to help identify people of interest by matching surveillance photos to known images — such as a headshot from a driver’s license or passport. But there are several barriers to its adoption, including high costs, unreliable results, and public opposition.”

In India too, the emerging AI technology for policing has faced opposition, with a growing chorus of critics demanding regulations. The Internet Freedom Foundation (IFF), a non-profit organisation that defends online freedom, privacy and innovation, says in its tracker that authorities have so far installed at least 48 facial recognition systems across the country.

“We at IFF does not support use of facial recognition by police/security agencies and has called for a ban on such use,” Anushka Jain, associate counsel (Transparency and Right to Information) IFF, told MoneyControl.

Well, law enforcement agencies deployed at the Kumbh fair are hardly going to take such a generous view.
Gulam Jeelani is a journalist with over 11 years of reporting experience. Based in New Delhi, he covers politics and governance for Moneycontrol.

stay updated

Get Daily News on your Browser