The Shadow War

On the Home Front: Scanners and Surveillance Systems

Sensors aren't merely for the battlefield; they will also play a critical role in protecting airports and passenger planes from terrorists in the United States. While low-tech measures such as stronger cockpit doors and armed sky marshals may make flying safer in the short term, the most promising long-term solutions are monitoring devices that will be almost invisible to airline passengers.

New scanning systems, for example, will be able to analyze everything in a passenger's bags, rather than just searching for specific objects. One cargo inspector, developed by the Silicon Valley company Ancore Corp., bombards luggage with packets of neutrons, producing gamma rays that paint a 3-D picture of the interior. The device also details the chemical elements present inside the package, making it easy to check for explosive materials.




Other scanners will be used for passengers at the gate. High on the acquisitions list are devices capable of detecting nonmetal objects that can be used as weapons, such as the plastic knives used in the September 11 hijackings. One such scanner works by passively detecting naturally occurring radiation as it reflects off an object, whose shape then appears on the operator's screen. Another option might be a handheld scanner, such as the one made by Massachusetts' CyTerra Corp., which uses ground-penetrating radar for the same purpose. Price is likely to be a consideration, though. An explosives detector called the Sentinel, which uses jets of air to knock trace chemicals off clothing, costs $125,000.




A high priority will be given to face recognition, a technology that has only been used experimentally at a few locations in the United States. This technology appeals to authorities because, unlike fingerprinting or eye scanning, it doesn't require the cooperation of the person being identified.




Face-recognition systems compare as many as 20 reference points on a face against a data bank of images, analyzing measurements such as the distance between a person's eyes. "Terror is not faceless," says Joseph J. Atick, chairman and CEO of Visionics Corp., a leading maker of face-recognition equipment. Existing systems, such as Visionics' FaceIt software, can find a match against 8 million images in a single second. However, face recognition may not work well in poor lighting conditions or crowds.




DARPA is working on an improved system called HumanID, which uses face recognition in combination with other surveillance technologies. The goal is to recognize individuals at distances up to 500 feet, even in poor lighting or weather. HumanID will also alert authorities when a person appears at one site on multiple days, or is seen at multiple locations within a facility. Such a system could help spot terrorists casing airports or other locations.



At the Los Alamos National Laboratory, researchers have developed surveillance camera systems that use neural networks to recognize human shapes and learn the normal patterns of behavior at a particular facility. When someone deviates from the norm -- for example, by waiting on a train platform but never boarding a train -- the system alerts security officers. Private companies are researching similar "smart cameras" to detect odd behaviors.



Recognition technologies are among the most promising defenses available for the home front. However, the effectiveness of any of these solutions may depend on the training given operators, which historically has been minimal. And if a lesson can be learned from the September
11 tragedy, it's that technology shouldn't be viewed as an impervious security blanket. Even if airports can be secured, there are other terrorist threats for which the United States is at present woefully unprepared. Foremost among them are biological and chemical attacks, which could cause the death of thousands. Prior to September 11, the idea of passenger jets being used to hit the World Trade Center and Pentagon was unimaginable. Afterward, nothing seems unthinkable.




Dan Carney, Mark Farmer, Michael Moyer, John R. Quain, Hank Schlesinger, and Bill Sweetman contributed to this report.