On 24th September 2015, a huge stampede caused the deaths of more than 2000 pilgrims who were suffocated or crushed to death during the annual Hajj pilgrimage in the holy city of Mecca. As the enormity of the tragedy dawned upon everyone, a familiar question did the rounds again, as it happens during every stampede. Could this stampede have been prevented? How can the next stampede be prevented?
In India, this assumes a bigger problem, due to the huge crowds that assemble during every major festival or pilgrimage. This makes the chances of stampedes extremely high. In 2005, over 300 people died in a stampede at the Mandiradevi temple in Maharashtra. In 2008, over 200 people died in a stampede at the Chamunda temple in Jodhpur. In the same year, over 145 people died in a stampede at the Naina Devi temple in Himachal Pradesh. This year, at the Godavari Maha Pushakaram in Andhra Pradesh, over 27 people died. As one can see, stampedes have continued to occur in India with alarming regularity.
Are there any lessons for the whole country from the numerous stampedes that have occurred, and can technology help in some way in preventing stampedes? One young innovative company quietly tucked in a small lane in Powai, believes it has the answer. The company, 2020 Imaging, has created what it calls ‘an effective alert management system’. Santosh Pillai, Director & CTO, 2020 Imaging, explains how the system works,” A stampede is visible to our naked eyes through situations like over-crowding, breaching of boundaries, and abnormal movement of the crowd either in direction or speed. What is visible to the naked eye is very likely to be caught through video analytics too. Once identified, appropriate response measures can be unleashed.”
Through an integrated alert management system, the company’s software identifies exceptions and triggers an alert for further analysis. For example, an alert could be given if the solution identifies sudden building of crowd from the cameras. An alert could also be given if a lot of traffic is getting accumulated around a particular traffic junction. In some cases, a citizen near the location clicks a picture of gathering crowd, and sends it to the command center from the 2020 App.
The solution also has the capability of identifying trends on social media, and can alert authorities if a protest is planned at a particular location and time. Depending on the alert sounded, an operator can quickly inform the necessary department and authorities to take further action. Santosh Pillai says that his company has perfected the technology of converging alerts — literally from every possible source.
Video analytics – a big game changer for the government
The firm believes that it has a big opportunity in India, as video analytics is part of the bigger story of visual intelligence. Governments and enterprises today exist in a dynamic setup governed by changing citizen behavior, consumer behavior, employee behavior, threat behavior and competition behavior. While data availability was the challenge of the last decade, data intelligence has become the challenge of the current decade.
“With video being more than half of all data that is generated in this world, video analytics is gaining quick ground. From cities to government establishments, from energy to transportation, from hotels to office spaces, from banks to retail – video-based intelligence is actively complementing and sometimes even substituting manual surveillance. In India, we have five clients who are using our integrated alert management platform along with video analytics. The spiritual institute of Global Vipassana Pagoda in Mumbai is one of our first clients and they are using our solution to not only identify exceptions but also manage and resolve them over a command and control infrastructure,” says Santosh Pillai.
Preventing terrorist attacks
Video analytics can also prove to be a big advantage in preventing terrorist attacks. Santosh Pillai cites the example of the Taj Hotel attack in Mumbai, where video analytics could have helped in curbing the impact of the attack.
“There are multiple things that could have been done, aided by video analytics. First, the suspicious person can be auto-tracked across multiple cameras and at any given point of time, we can know his last location. Second, the number of people inside can be tracked floor-wise and zone-wise. Third, the evacuation mechanism could be visually enabled by accurately tracking the crowd on the floor maps. Fourth, the command station system can quickly talk to other sub-systems like access control, elevator control, public addressing system, etc. and enable integrated response. All these could have saved precious time and lives,” states Santosh Pillai.
The success of the solution was proved when the system averted a potential threat scenario for one of 2020’s clients. During a VVIP visit, though the premise was cordoned off, a wrong parking of a two-wheeler at the last minute was caught on the 2020 video analytics system. This alert set off a series of actions including authority notification, vehicle inspection and removal. This was completely managed on the 2020 system. At another site, a prolonged loitering by an individual during off-hours around a water-facing boundary of the facility was detected on the cameras. The system auto-alerted the officials on duty on their handhelds. Upon investigation, it was found that the individual was planning a suicide.
Given the usefulness of the solution, governments have shown keen interest in deploying this solution. Today, 2020 is working with one of the state governments to develop a central command station connecting multiple cities across the state. The solution will deploy some of the cutting-edge video analytics functionalities such as facial recognition, number plate recognition, trespassing detection, mob gathering detection, abandoned baggage detection and even detection of vandalism attempts like property destruction and camera tampering. Additionally, the firm is also providing a complete collaboration and communication framework along with big data integration, enabling proactive actions to situations.
Today, video analytics is being actively explored for business intelligence and analytics. From governments ,hotels and automotive companies, video analytics can be used everywhere.
Santosh Pillai cites some examples, “A toll management company can use video analytics for vehicle count and classification and ensure there is no slip-up by the collection agencies. A retail store can track the footfall and crowd movement, real time. An automotive plant can integrate cameras with the production system and identify anomalies in the assembly line. A power plant can use cameras to identify O&M gaps through visual auto-inspection of the production equipment. A hotel can use cameras to detect queues in the lobby as well as track the pattern of usage of different services in the facility. Laboratories, for example, can use specimen movements to generate and log alerts. e-Commerce platforms can use video-based intelligence to predict returned goods based on differences between sample images and actual looks. The potential is unlimited, when it comes to video-based intelligence.”
Managing cities by detecting exceptions
Given the fact that India is now on an overdrive for building smart cities, the opportunities for smart cities are immense. Cameras, sensors, energy, water and telecom systems, can all be brought together on to one platform.
“From deploying video analytics to integrating social media and big data analytics, the city can be managed by exception. For example, facial recognition based on police database can detect criminals in the city with reasonable accuracy. Traffic violations can be captured along with automatic number plate recognition (ANPR) and dynamic penalty system can be imposed through integration with RTO database. Police, Fire, Ambulance and Traffic resources can be viewed on the maps, real time, and deployed from the nearest locations,” states Santosh Pillai.
Citizens can become part of the e-governance process as they can just click pictures and videos of anomalies and these can reach the control room with exact latitude and longitude along with feed from nearest surveillance camera.
Given the huge potential of video analytics, its usage is only limited to the imagination of the enterprise which deploys it. Deployed effectively, video analytics can truly be a big game changer for the entire country.