Emerging Technologies: Computers Watching Over Us

Video Analytics -- software that automatically monitors video feeds from surveillance cameras -- can identify certain types of homeland security threats. And the technology is getting better all the time.

by / December 20, 2006
With cameras installed at every turn monitoring a person's activity, Big Brother was a frightening concept in George Orwell's classic novel, 1984, where citizens of a fictitious totalitarian government were watched and punished for slight indiscretions. What many people today may not realize is that Orwell's vision has become more real than they thought and they are indeed being monitored throughout the day as they work, shop, bank and drive. But that rather than a tool to utterly control their lives, these cameras are helping to keep them safer.

Some estimates put around 30 million video surveillance cameras in the United States, shooting four billion hours of footage each week. The video surveillance industry has grown into a $160 billion global industry, especially after homeland security efforts after 9/11, where the federal government has poured money into video surveillance.

Numerous cities have set up cameras in various locations many increasingly connected with wireless technologies, with traffic lights only the beginning. New York City, with the help from a Department of Homeland Security grant, has created a network of thousands of cameras throughout the city, including subway stations, traffic signals and private businesses. The New York City Police Department operates its own network of 3,000 cameras. Chicago has also received grant funds from the DHS and built a "Homeland Security Grid" of 2,250 cameras, with plans to add even more cameras in the coming years. Baltimore and New Orleans also have thousands of cameras throughout the cities.

Municipalities that deploy Wi-Fi networks can often utilize the infrastructure to greatly reduce the cost of deploying video cameras across a city. And emerging mobile video capabilities are going to play an increasing role for police and emergency responders. So there is little doubt that the number of video cameras is going to increase dramatically in over the next decade.

Staffing the feeds from thousands of cameras would be a budgetary and logistical nightmare, yet with video analytics software that tracks potential problems and monitors areas for preset situations, the problem may be solved.

Video Analytics is a software system at the forefront of security technology that is helping to minimize the need for human viewers to distinguish important events from video feeds. Video analytics software tracks live video monitor feeds and pinpoints video images that fit specified criteria. When the software identifies predetermined criteria, such as a person loitering, or a bag left at the airport, or a car in a restricted area, it will then set off appropriate alarms to security personnel.

"Traditional video services are very reactive and you have to make sure to pay attention to cameras and try to make sense of what's going on, where with video analytics, the software uses mathematical algorithms that actually shift through all the videos and triggers alerts if something goes wrong," said Dilip Sarangan, research analyst for Frost and Sullivan.

Video analytics software groups video pixels into objects onto a computer database, which are then compared with present behavioral and motion parameters. If a preset object or motion parameter is detected an alarm will be triggered.

Security systems have benefited the most from the emerging technology, since the technology helps identify problems and notify appropriate security personal. The largest user of video analytics currently is federal and state governments, taking up more than half of the market share, according to Video Analytics: The Ground Reality, a report on video analytics by Frost and Sullivan. The Department of Homeland Security utilizes the technology and has offered grants to state and local authorities to purchase and deploy video surveillance systems for homeland security applications. Under Homeland Security, video analytic software has been mainly used for monitoring security at airports, shipping ports and borders.

One of the biggest users of video analytics are border protection officers, who use the software to detect things like the breaching of a perimeter fence, or spotting of people in border areas, according to March Networks, a company that provides video analytic software solutions. Critical facilities such as nuclear power plants, military installations and telecommunications hubs have a higher level of protection through video surveillance.

Video Analytics also assist in all aspects of transportation safety, including security at airports, highways and bus and train services. The software can identify and send alerts for suspicious bags and packages, people accessing restricted areas, the size of a crowd and abandoned vehicles. The technology can also help to count people and vehicles in an area at any given time and safety hazards can be spotted.

The second largest market for video analytics software at the moment is retail, banking and gaming industries, according to Frost and Sullivan. Banks use video analytics for security at ATM machines and inside banks, while casinos use video analytics for security and to spot trends in gaming.

Retail companies use video analytics to alert security staff to suspicious transactions such as check authorizations with no manager present. Also, security can be alerted to suspicious loitering and unauthorized access to storage areas. Another feature retail businesses have been taking advantage of is the ability to detect and program patterns of shoplifting, with video analytics able to serve as a type of search engine to filter through tape data to find shoplifting evidence. Store managers can monitor cash register lines and store traffic, while marketing specialists can obtain statistics.

Corporate campuses and factory plants also benefit from additional video intelligence, by being alerted to unusual after-hours activity, trespassing and vehicles in prohibited areas.

While video analytics is a relatively new technology that holds a lot of promise, it is expected to be further developed in the coming years. The biggest interest in video analytics has been from government organizations interested in the ability of video analytics to recognize license plates and faces, Sarangan said. However, the problem now with this form of video analytics is that it works well in a controlled environment when a picture is clear, but in a natural setting video analytics can be unreliable.
"The idea is a car drives by a school and a camera catches the license plate, which is then sent to a DMV database and we know exactly where they are," Sarangan said. "In a very controlled environment it works, but in high speeds it depends on the angle and how the picture is taken."

Another obstacle facing video analytic software is that it has difficulties tracking video images that display constant motion. Another impediment to the widespread use of video analytics technology is the cost of digital camera system upgrades, since only high quality cameras and lenses will provide adequate video stream for accurate video analytic results. And with the additional costs of software and computers, many organizations have not yet investing in the new technology.

Yet many companies in the security and video monitoring industry feel these problems will be ironed out in the coming years. General Electric has invested heavily in the global security industry and has researchers creating video surveillance systems that can detect explosives by recognizing electromagnetic waves given off by objects. GE is also developing programs that can identify distress in a crowd by honing in on erratic body movements.

The first totally geo-spatial video surveillance system, called GView, was recently made available by Guardian Solutions. The GView technology upgrades a digital video recording surveillance system to a 3D situation system connected with Google Earth for rapid threat detection. As the camera records events, GView isolates and tracks threats, which can be an individual, vehicle, or vessel and uses Google Earth to display all movements three dimensionally.

IBM is developing the "IBM smart surveillance system", which promises not just the ability to automatically monitor a scene, but also the capability to manage the surveillance data, perform event based retrieval, receive real-time event alerts through standard web infrastructure and extract long term statistical patterns of activity.

The fact is technology often outpaces society and video analytics promises to be a big part of security as society catches up. As the number of video security systems grows, the need for computer video analysis also will expand. By 2012 it is expected that retail, banking, gaming, corporate facilities, and transportation industries will take over the majority market of video analytics technology. If so, when that time comes, a form of Big Brother -- or rather a lot of Little Brothers -- will indeed be monitoring our every activity. But if it helps to keep us safer, it seems people for the most part won't mind.

Chandler Harris is a contributing writer for Digital Communities.

Chandler Harris Contributing Writer