Organisations world over are increasingly using video analytics software to identify people, vehicles, and monitoring dangerous environment and situations
Pratap Vikram Singh | August 24, 2015
Closed circuit television (CCTV) cameras are everywhere. They are being used in restaurants and coffee shops to monitor kitchen hygiene and customers’ belongings. The oil and gas companies use them to ensure security of gas and oil pipelines. The security and police forces rely on a network of cameras connected to a control room or a command and control centre to guarantee safety and security of human lives and physical assets in cities, airports, and railway stations. Of late, working parents have resorted to hidden surveillance to keep an eye on babysitters.
The technology related to surveillance has evolved in the last two decades. Organisations world over are increasingly using video analytics software which detects or identifies a person or a vehicle or an activity from a live camera feed. Consider the case of software giant Microsoft. It has offices at 800 locations across 200 countries. The company has deployed 23,000 cameras across these locations to ensure safety and security of its 1,15,000 employees and physical assets. The video feed from these cameras is monitored from two operation centres: one in Redmond, Washington, and another in Hyderabad.
Even more interesting is the fact that teams consisting of only five people oversee huge video walls in these two cities. Instead of recruiting more people for the job, the multi-billion dollar firm has developed its own analytics software to monitor the video in real-time and sift petabytes of data being generated by thousands of cameras twenty-four-by-seven. [The details of the analytics software are not in the public domain. When contacted through email, Brian Tuskan, senior global security director of technology and investigations, says, “We are working on some new technology and I am not at the liberty to discuss at this time.”]
Video analytics applications, in general, either come embedded with CCTV cameras or installed in servers. Using analytics, agencies can detect, among other things, unattended objects, a crowd gathering, identify and track people and vehicles, rationalise signal timings at traffic junctions, and raise alarm in case of an intruder. It is an integral part of anti-crime and terror strategy in the US and the Europe.
How the pioneers are doing it
The American and European cities are pioneers in using surveillance and analytics. London and its suburbs alone have over 2,00,000 cameras, most of them deployed by private users, said Subhash Patil, partner, PricewaterhouseCoopers (PwC). Some reports estimate that there is one camera for every 11 people in the United Kingdom. Washington DC and New York are among cities under wider surveillance coverage. The police department in these cities uses analytics to crack down on local crime and prevent terror attacks.
The metropolitan police department of Washington DC uses a visualisation platform, developed by Tokyo-based Hitachi Data Systems, to collect data from public and private cameras, mobile phones, and social media. The solution correlates disparate data and video systems geospatially. In case of an eventuality, the platform will provide situational awareness to the agencies, Vivekanand Venugopal, vice president and general manager, Asia Pacific, HDS, told Governance Now.
While Indian agencies have been using video footages to solve criminal cases for over a decade, they have just started procurement and use of analytics. Surat police is considered a pioneer in use of video analytics for the maintenance of law and order. The city police has deployed NEC’s analytics tool, including facial recognition software, which helps it in tracking the movement of criminals in the city. The city police has uploaded images of 20,000 criminals in the system. “Whenever these faces appear in camera frames, the system generates an alert. The application has shown an accuracy of 75 to 80 percent,” said Rakesh Asthana, police commissioner, Surat police.
They have formed a picture intelligence unit (PIU), which uses facial recognition and automated number plate recognition system for policing. The software used by PIU enhances the quality of images, searches against mugshot databases, and locates potential suspects. Other analytics being used by the Surat police includes unattended baggage, crowd gathering, automated number plate recognition and speed detection. The city is under surveillance of over 600 cameras across 113 strategic locations from the police point of view. Eventually the city police plans to deploy 5,000 cameras, out of which 200 will be connected with video analytics.
A major surveillance project, said Patil of PwC, which can set standard for rest of the city police, is being rolled out in Mumbai. Costing close to Rs 950 crore, the project includes deployment of 6,000 cameras and advanced analytics aimed to equip the police to deal with possible terror attacks. Using different analytics software, Mumbai police will be able to detect a gunshot, unidentified object, breach in perimeter security, vandalism and tempering with CCTV camera, crowd gathering, particular individuals or vehicles, illegal parking and over-speeding.
The cameras deployed are of 2 megapixels and capture 15 frames per second – good enough for live streaming. It is important to note here that these cameras use latest compression technology (H.264 compression technology) to send video over the network, and hence would lead to substantial savings in terms of storage and bandwidth storage. Even the compressed round the clock video feed from 6,000 cameras (even if it is 1 megapixel resolution) would require 25Gb storage space. For 30 days, it will require 15 petabytes of storage space. No doubt the storage component itself has a cost of approximately Rs 35 crore.
Being executed by Larsen and Toubro, the surveillance project follows a collaborative monitoring model, also being followed by the city of London. In case a situation similar to 26/11 emerges, said Patil of PwC, collaborative monitoring will allow agencies to plug in privately-owned cameras deployed in hotel, colleges, institutions and commercial centres to the central command and control centre of the Mumbai police.
The system has been designed such that 102 institutions – including NSE, BSE, mantralaya, raj bhawan and schools, among others – will be connected to the command centre so that cops could monitor and analyse the video feed in case of an incident. If there is any feed coming from any other city, the police will be able to use and analyse it from the command room. The overall analytics solution includes basic video management, audio and video analytics and a GIS platform. The project is being implemented in three phases; the rollout will be over by the end of next year. A similar project for Delhi has been on the anvil but it is awaiting approval of ministry of home affairs.
Indore police has deployed traffic incident violation detection developed by Kolkata-based Videonetics. Important locations and 15 traffic junctions have been covered by 200 high definition IP CCTV cameras. The feed, which is carried to the control room using optical fiber, is analysed for violation. The red light violation is detected in real time. The license plate is recorded and a challan is issued based on the details fetched through Vahan database of the transport department. The system captures multiple snapshots which can be used as judicial evidence.
The project implementation started in December and it was over by June this year. [Till first week of June, system had generated 36,000 challans. Out of this, 8,000 challans have been paid by defaulters and an amount of Rs 40 lakh has already been received. For law enforcement the police plans to deploy 400 cameras]. Banks are major consumers of surveillance and analytics. At present the cameras fitted with the ATM machines store data locally. The video is recorded and is not transmitted to any control or operation centre for real time monitoring. Analog Devices, a US-based firm, has devised an analytics and image processing box, which is integrated to the ATM’s camera. It sends an alert to the agency in charge of the security in several cases: multiple people entering a kiosk, a scuffle, a person brandishing a weapon or falling down, getting shot, injured or sick or a person with a mask or helmet on. At present the solution is being demonstrated and piloted with system integrators, said managing director of Analog Devices Somshubhro Pal Choudhury.
The 73 airports under the airport authority of India (AAI) are also using advanced surveillance and analytics, said a general manager level official with AAI. The solution includes video management, and analytics to detect unattended object, zone detection no entry areas. Usage of surveillance and analytics, however, is highly contested by the security agencies. Security officials said that even though the cameras have inbuilt analytics it is never used. “Neither Delhi or Mumbai airport, having over 4,000 cameras each, and teams of around 15 people watching video walls, use video analytics,” confirmed a serving assistant inspector general of police rank official working with central industrial reserve police force (CISF). The force guards all airports, government premises and critical facilities across the country.
Another major challenge is the lack of coordination between the agency making the procurement of surveillance cameras and the security agency. Here is an incident which elucidates this gap. The Delhi airport in total has over 4,000 CCTV cameras of varying specifications. A team of 15 CISF officials man the control room fitted with huge video walls. The video walls displayed videos from these cameras one by one and it was manually monitored by the force officials. In a multi-stakeholder workshop, said senior officials related to airport security, the CISF officials were told by the AAI that the cameras in use at the airport had inbuilt analytics, which could be used to identify unattended objects or for facial recognition. The situation is similar at the metro stations, officials said. Delhi metro rail corporation uses thousands of Bosch pan-tilt-zoom cameras, which come with built-in analytics, to monitor over 140 stations. The analytics feature is rarely used.
Yet, another challenge is the maintenance of deployed cameras. When agencies procure solutions for video surveillance they are also expected to buy annual maintenance. During the review visit to the Delhi airport, officials reviewing the security found that 20 to 30 percent of the cameras were not operational. It was primarily because there was no annual maintenance contract. “If this is the situation at the Delhi airport, imagine how things would be at other airports,” an official said.
Cameras being used elsewhere in the capital were procured five years ago during the commonwealth games and after 26/11. In absence of maintenance, the cameras are non functional even in prime locations such as Connaught Place, said Virag Gupta, a Delhi resident and a senior advocate with supreme court. Even the functional cameras, said Muktesh Chander, special CP, traffic, Delhi police, are of low resolution. “You can’t run video analytics on these cameras,” added Chander.
Another major challenge, said Arvind Ranjan, former CISF chief, is that the video analytics needs to evolve and it needs a lot of customisation. “In the field of security there is nothing readymade. The solution has to be tailor-made as per the requirements,” added Ranjan. “The lack of coordination makes things worse.” Consider the procurement of perimeter intrusion prevention system by AAI. The agency procured the solution from an Israeli company for Rs 38 crore as it wanted to cut down the number of CISF personnel — who are paid by the user agency (in this case AAI) – deployed at the airport. While reviewing the solution, the CISF personnel found that video surveillance coverage was inadequate and hence was of little use for security. “In a day the system generated 500 to 600 alarms. On every alarm the CISF personnel had to visit and inspect. Most of these alarms were found superficial,” a senior official told Governance Now. As a result the CISF refused to rely on the perimeter intrusion software and cut down on the number of personnel at the airport. “Capacity building of security officials – so as to acquaint them with latest technology – is yet another challenge,” said Ranjan.
While surveillance cameras and analytics, when operational and in use, strengthen security, it poses a challenge to individual privacy if used in an unregulated manner. To start with, India doesn’t have any standard operating procedure or a rule book defining why and how it can be used, says Sunil Abraham, executive director, centre for internet and society. The European Union and the US follow regulations. United Kingdom follows basic principles which includes 1) notice, stating who installed camera, how to communicate displeasure; 2) right to access which entitles people, the subjects of surveillance, to get a copy of the footage; and 3) transparency about the practice. He said that Justice AP Shah led committee on privacy in 2012 proposed nine principles safeguarding individual privacy: notice, choice and consent, collection of limitation, purpose of limitation, access and correction, disclosure of information, security, openness and accountability.
Experts believe that a better coordination between agencies developing or deploying a solution and agency handling security could go a long way in strengthening security with the help of technology. More number of cameras should be deployed at critical facilities and public places. Capacity building of security personnel at regular intervals is of utmost importance to ensure utilisation of technology. Besides, the government must draw a rule book for operating CCTV cameras and safeguard individual privacy.
PM Narendra Modi’s yet another niftily acronymed scheme, UDAN – short for Ude ‘Desh Ka Aam Naagrik’ and otherwise called ‘Regional Connectivity Scheme’ in officialese – got off to a flying start on Thursday. Modi formally launched a flight from Shimla to Delhi, and
He accompanied his father to film studios in Chennai and helped him in designing sets, but Thota Tharrani wanted to be an artist. So he studied mural painting and print-making, but as luck would have it, he finally returned to tinsel town. And the world soon took note. In Mani Ratnam’s pa
Is the AAP headed for a split?
A sale-purchase agreement was signed between Numaligarh Refinery Limited (NRL) and Bangladesh Petroleum Corporation (BPC) for supply of high speed diesel (HSD) through the proposed 131 km Indo-Bangla friendship pipeline. The agree
The dismal performance of the Congress in the Municipal Corporations of Delhi elections forced party chief Ajay Maken to announce his resignation, ending an energetic effort to revive the party in the national capital. Ajay Maken, now 53, had taken over as the chief
The BJP’s clean sweep is not just a referendum on the Arvind Kejriwal government, but also could mark the beginning of the end of one of India’s youngest political parties, the Aam Aadmi Party (AAP). After the Bharatiya Janata Party’s massive win in the UP assembly elections, th