The departments of Defense and Homeland Security have invested in technology to prevent attacks like the one in Brussels, including facial recognition technology that can spot and flag a suspected terrorist in a car heading toward an airport or crowded subway. But bringing that technology out of the lab and getting it to airports and street corners is a lot harder than just snapping a photo.
Just hours after two bombs exploded in the public departure area of Zaventem airport and a third at the Maelbeek metro station, grainy closed-circuit television footage surfaced showing three men at the airport. Two wore black gloves on their left hands, the third a white coat. Security experts said that footage showed the likely attackers and that the gloves might have concealed detonators.
The attack on Zaventem doesn’t point to a specific weakness in airport security, said one DHS official who spoke anonymously because they were not authorized to discuss the incident with media. “The reality is, they did it in a public area of the airport. The assumption that they determined that it would be too difficult to get any deeper into the airport is a reasonable assumption to make,” the official said. “When events like this happen, we play extraordinarily close attention to them. Could it happen here? If so, how could we intradict it?”
In order to thwart an attack, a facial-recognition-based system must have access to a database that already contains the would-be perpetrator’s face, sensors that can obtain usable snapshots of people approaching the protected area, and a way to alert guards or otherwise cut off access to the target. Getting all three of those at once is the challenge of securing a place like an airport’s departure area or a metro station.
In 2014, the military tested a General Electric high-speed, multi-resolution camera capable of capturing a facial image even at an angle. The system was designed to ID someone in a moving car headed toward a base, but could be deployed on city streets or on the road to an airport.
“The high-resolution image can be panned and tilted using an internal device and the system is designed to capture images at a standoff range of 22 to 33 meters. That’s up to 108 feet,” John Boyd, director of Defense Biometrics and Forensics, said in 2014. “The … images are then outputted to a machine-vision algorithm to continuously track the location of the vehicle. The vehicle coordinates are then used to estimate the time of arrival of the vehicle to four fixed catcher zones along a calibrated track. Some 20 high-resolution images are captured per zone as the vehicle travels five miles per hour — five to ten. The high-resolution frames are then automatically analyzed and match scores are fused into an average score to positively identify the individuals in the vehicle; and this is with tinted glass, among other things.”
In tests at Fort Belvoir, the system worked, but slowly, a Defense Department official said at the time .
“It was a demonstration sort of environment, so the van would make its approach but then have to slow down so that the computer, which was not optimized” could ingest the data. The official called the system “not deployable” but “a very positive step.” He said, “We were able to identify people in a moving vehicle. There is the feasibility of the technology being further advanced and then point it at a normal post.”
Big brother or safety? There is a trade-off. Which would you opt for? Seems that this type of terrorism is not going away anytime soon.