Millimeter-wave imaging speeds weapon detection

Conventional security systems include metal detectors for personnel and x-ray systems for hand-carried items.

Conventional security systems include metal detectors for personnel and x-ray systems for hand-carried items. To detect plastic or ceramic handguns and knives, plastic weapons, and liquid explosives, conventional metal detectors and x-ray systems can be augmented with millimeter-wave imaging that penetrates clothing but is reflected by the body and concealed items.

Such a system, now in prototype stage, has been developed by David Sheen, Douglas McMakin, and Thomas Hall of the Pacific Northwest National Laboratory (PNNL; Richland, WA). Consisting of a vertical millimeter-wave array of antennas scanned in a cylindrical manner about the person under surveillance, images from the system reveal differences in shape and reflectivity and allow operators to readily identify concealed weapons.

In operation, the person scanned stands on a platform near the center of the scanner and is illuminated by a wide-bandwidth, diverging-beam coherent millimeter-wave source. Simultaneously, a millimeter-wave array with 384 elements spaced over a length of 1.92 m is rotated about the person. Operating over a 26- to 30-GHz frequency range, the system can produce resolutions of approximately 1 cm laterally by 3.75 cm in depth. Reflected signals are then recorded coherently by a receiver, digitized, and stored in a computer. These data are then mathematically reconstructed to form a series of three-dimensional images of the target from various illumination angles. Using a back-projection algorithm similar to those used in medical CT systems, image reconstruction relies on the use of the fast Fourier transform.

A digital-signal processing (DSP) system using eight Alacron (Nashua, NH) boards, each containing four Analog Devices SHARC DSP processors and 128 Mbytes of RAM reconstructs the data. The DSP system is capable of reconstructing eight output frames in approximately 10 s. A 3-D combined cylindrical reconstruction algorithm has been developed that uses similar cylindrical holographic-imaging techniques to the new prototype. This technique combines the results of multiple 3-D cylindrical image reconstructions to form a combined image, which contains information from every illumination angle.

The combined 3-D image is rendered by projecting through the data set at discrete angles over the full 360° angular range. Imaging studies have been performed with this new technique using a laboratory scanner and wide-bandwidth millimeter-wave transceivers on mannequins. The 3-D combined holographic imaging data provide a full volumetric data set where every pixel has data obtained from all illumination angles. This dramatically reduces shadowing that is present in the original cylindrical imaging technique.

For the prototype cylindrical system, which operates from 26 to 30 GHz, a bandwidth of 4 GHz is sufficient for good depth of focus. Lateral resolution is approximately 1.0 cm, whereas the depth resolution is 3.75 cm.

"Despite the overwhelming success of results from the system, public acceptance of this type of imaging may be a significant hurdle to implementation of this technology," says McMakin, program manager. "Although the images are low resolution by optical standards, they do reveal anatomical characteristics of the person being screened," he says. At present, the US Federal Aviation Authority is opposed to presenting this type of image directly to an operator, and PNNL is investigating pattern-recognition and image-segmentation techniques to eliminate operator intervention.

More in Home