A Smarter Image: New digital systems can see all, but they dont tell all
Jul 22, 2002 12:00 PM
In the 1995 attack on the Federal Building in Oklahoma City, Timothy McVeigh parked a truck in front of the building and walked away. After the explosion, investigators reviewed videotape from a camera watching the front of the building and identified the truck as the weapon used in the attack. No one noticed the truck before the explosion.
The reason: Conventional video monitoring systems report too many facts for human monitors to evaluate and act upon. Security officers monitoring comprehensive CCTV systems do not always notice and act on potential problems.
Today, a new generation of CCTV technology is addressing this shortcoming with systems that can sense potential security problems and alert security staff in time to respond.
These smart video systems employ digital technology that treats video as data composed of bits and bytes. To locate the video on a disk drive and replay it in seconds, a security officer simply types a date and time into the system.
Conventional CCTV cannot do this because it uses analog technology. A computer can analyze digital data stored as bits and bytes, but not analog information, which is composed of electrical signals "analogous" to the video seen by analog cameras and stored on videotape. As a result, finding an appropriate clip of video tape takes a long time. The tape from the camera that recorded the event must be located and plugged into a VCR. Security officers must fast-forward through what might be hours of images to find the incident in question.
Smart digital video overcomes this problem with instant retrieval. More importantly, smart video systems can evaluate video. Just as you can set a computer software application to monitor data about a project's cost stream and to alert an impending cost overrun, security personnel can set digital CCTV systems to monitor the front entrance of a building and call attention to a vehicle that has been parked more than a few minutes. Smart video systems can monitor floor space in a building or airport lobby and set off an alarm if a package or bag has been abandoned for more than a few minutes. They can alert security officers if a person enters an access-controlled area without authorization or enters a building through a door and corridor intended for exiting.
The software behind these systems can also automatically recall video from a few minutes before these incidents occurred and capture an image of the person who left the vehicle or the package, allowing security personnel to circulate the photo to patrolling officers over networked handheld computers. Additionally, these systems can record images of individuals checking in at an airport ticket counter and compare those images with an image of the person at the airline gate to determine if the person boarding the plane is the same person who checked in.
Digital video technology solves the problems associated with analog technology and brings a host of new capabilities to CCTV security.
"This technology makes security proactive through advanced video content analysis," says Jonathan Moav, director of marketing for NICE Systems, an Israeli digital video company with U.S. headquarters in Rutherford, N.J.
Moav cautions, however, that not all digital video systems have these capabilities. "There are three levels of systems," he says. "The first level is basic and simply replaces the VCR with a disk drive as the recording device. The second level allows searching by date and time and can accommodate more cameras."
The third level, Moav says, will accommodate virtually unlimited cameras by linking systems together. In addition, security officers can ask these systems to conduct swift searches for faces and specific shapes. Third level systems can also alert to defined events: a person moving the wrong way through a corridor; a bag abandoned for more than five minutes; or a vehicle parked longer than two minutes in a restricted area. These systems make it easy to distribute video clips showing faces, bags or vehicles that need investigating via e-mail to security officers with handheld computers. "This is real time analysis, at a level critical to government installations," says Moav.
How is it possible?
"We've written software that can identify objects," explains Larry Bowe, director of product management and strategic alliances for Loronix Information Systems of Durango, Colo. "The software allows you to set parameters for shape, size, direction, and speed. For example, you can tell the system to notice a truck larger than a certain size that has not moved for a certain period of time."
In one government facility, for example, a Loronix digital video system works with a metal detector, X-ray scanner and card access system. "When people card in for the day, the system takes their picture along with any belongings sent through the X-ray machine. The software associates all the information with the person," Bowe says.
If a bag is later found abandoned, the system can instantly retrieve a picture of the owner.
Sophisticated smart video systems require complex planning to implement, says Jim Francis, senior vice president, security services group with Kroll Inc., New York. "You need a well-engineered plan for deployment, qualified installation and training that goes beyond how to use the system," he says.
Engineering errors can limit the use of a smart video system, continues Francis. Suppose a security manager wants to retain digital images in the system for 30 days. Misjudging the system specifications required for that amount of storage can produce a system with only 10 days of storage and compromise performance. "The engineers must make a complete appraisal of issues related to the system and establish storage and retrieval capabilities strong enough to support the system's goals," Francis adds.
Digital video systems must also allow for adequate expansion, Moav says, cautioning against trying to expand by adding a second system incompatible with the first. "The ability to work on a network is important, too," he says. "Networked systems can be contacted from anywhere if you have a password. They also allow those with passwords to tap into video within an installation from anywhere. If a building is attacked, security officers outside can find out what is going on inside."
When it comes to installation, Francis recommends stringent review of possible integrators to ensure the chosen company has capabilities matched to the sophistication of the digital video system.
Poor training can compromise the effectiveness of the best designs, Francis notes. "The biggest mistakes I've seen with these systems involves insufficient training for operators," he says. "It is a mistake to assume that the integrator will provide a full complement of training, without direction from the owner about the facility's security needs. You need to make sure that you train people about more than just how to push the system's buttons. The security staff must have training related to the security concept behind the system."
According to Moav, manufacturers will continue to improve the performance of digital video systems through software improvements.
For example, today's systems can identify faces under controlled conditions ¡ª if the lighting is adequate and the subject is looking at the camera. At airport ticketing counters, security officers can control lighting and the subject's position in relation to the camera. Similar control is available at boarding stations. In time, smart digital video systems will grow smarter.
Tomorrow's systems will be able to pick faces form a crowd moving through a poorly lighted hallway.