Â鶹´«Ã½

Security screening at airports and railway stations could be streamlined using a new high-frequency radio scanner that automatically spots concealed weapons.

The scanner detects millimetre-wavelength radio waves reflected off a target and converts the reflected signal into a 2D image. Computer software performs a rapid analysis of the scanner imagery, searching for incriminating objects.

Millimetre-wave radiation passes through clothing easily and is absorbed and reflected by other materials to different degrees. This makes it well suited to security screening (see All-seeing scan spares your blushes).

It can be used to spot hidden handguns and knives, and even non-metal weapons such as ceramic knives. Until now, however, a human operator has been needed to pick out concealed weaponry.

Rapid screening

The new system could therefore speed-up routine security checks by rapidly screening for hidden weapons, says developer Yvan Petillot, a computer vision researcher at Heriot-Watt University in the UK.

In tests, participants were asked to stand a few metres in front of a scanner developed by UK firm QinetiQ and turn through 360°. Some were also given a small handgun to hide beneath their clothes. The computer software was able to identify the gun based on its shape and track it across several successive frames using a statistical image processing technique.

Research team-member Emanuele Trucco says the system could have benefits over conventional airport metal detectors. “This can go further because it can say exactly what and where a metal object is without the need for a touch or strip search,” he says. Trucco also suggests that a computer scanner could be more reliable than some human operators.

Sharper images

The team also used the system to identify non-metallic weapons, although it was less reliable for this. “We can detect metallic objects very well,” Petillot says. “But ceramic knives are much harder.”

Petillot adds that newer high-frequency scanners could sharpen images and work at greater distances: “The hardware is advancing so I think that in around two years it could be possible to spot [ceramic knives] automatically, like we can with metal objects.”

Graham Smith, who works on millimetre-wavelength imaging at St Andrews University in the UK, says the approach shows promise. “It is amazing what you can see with millimetre wave,” he told Â鶹´«Ã½ (see graphic).

But Smith says the Heriot-Watt University system will need to be fine-tuned to meet with stringent security screening standards. “I’m not sure they are specific enough to be as reliable as a metal detector yet,” he says.