Any way you look at it, video surveillance technology is becoming more sophisticated.

A recent report by ABI Research projects that revenue generated from surveillance software will more than triple from US$245 million this year to US$900 million-plus in 2013. Technological breakthroughs have added several more capabilities to surveillance cameras in recent years, from face-recognition software to cameras designed to mesh with radio-frequency identification tags. Added to this, ABI Vice President and Research Director Stan Schatt says that the advent of Wi-Fi has made it possible to place wireless cameras just about anywhere while still being able to send footage back to a central location.

"More and more governments can use these cameras outdoors in places where they couldn't before," he says. "This is because they can use Wi-Fi to send signals back to the security department, whereas before they couldn't get cameras properly hooked up to the network."

What modern surveillance can and can't do

IBM has become one of the more prominent players in the advanced video surveillance market, as the company in recent years has helped Chicago roll out equipment and software for its Operation Virtual Shield crime detection and prevention program. And while many governments and businesses are using IBM surveillance systems for traditional security operations, IBM physical security CTO Arun Hampapur says that new surveillance technology is opening avenues that were not possible 10 years ago.

For example, Hampapur says that you can program a camera to watch a perimeter around an airport where people are not allowed to pass through. So if a person jumps a fence or tries to access the area without being authorized, the camera will automatically detect his presence and send out an alert to the security department.

Ed Troha, the director of global marketing for surveillance software vendor ObjectVideo, says that his company has developed techniques for analyzing pixels and groups of pixels by the way they interact with each other in motion video. In other words, the software not only analyze still pictures but also looks at how different objects are moving in relation to each other in real time.

"What intelligent video endeavors try to do is to detect, track and classify objects within the video," he says. "So intelligent video can tell you when there's a person in an area that is designed only for vehicles. That's a very good application of intelligent video, and there's a distinction between this kind of technology and technologies such as facial recognition software. Those technologies analyze pixels within the video, but don't analyze the scene itself."

Hampapur says that this type of technology has evolved to the point where cameras can detect, index and catalog the movements of objects and categorize them by their size, color and shape. Thus, if police are looking for a white van that has been used in a series of robberies, surveillance cameras will be able to spot and index all white vans that pass through their line of vision and send that data back to the police department. Hampapur notes, however, that the technology has not advanced to the point yet where cameras can detect individual faces of people who have outstanding arrest warrants in their names. Face-recognition software, he says, still requires active compliance from the individuals being scanned - that is, people trying to enter certain areas must look directly into the camera and hold their faces still until the software has a chance to scan and authorize their face for access.