Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS

NOVA Online (click here for NOVA home)
Killer's Trail
Site Map


Diagram of interference pattern generation A diagram showing the splitting of the laser light into two beams, which results in the fringes that lie at the heart of AFI.
3-D Mug Shot
Part 2 | Back to Part 1

So how does AFI work?
First, a laser illuminates the subject with a pattern of stripes, or fringes. The pattern comes from splitting the laser beam into two point sources of near-infrared light, each bearing a wavelength of 780 nanometers (780 billionths of a meter). By making fine adjustments to the two sources of light, operators can manipulate the thickness and placement of the fringes - hence the term Accordion Fringe Interferometry. (Interferometry is the measuring of objects using interference patterns as a tool.) To take accurate readings of complex shapes, such as a nose, operators must be able to adjust the fringes. Initially, they lay only a few thick fringes over the subject, allowing a reading of the general outline of the head. When they thin the fringes to get a more accurate reading, that understanding of the general shape becomes crucial in order to fill in the details - seeing the forest despite the trees, you might say.

Imagine stripes of light streaming through a venetian blind. Hold a ball up to the light, and the stripes curve around it. If you were to observe the ball from the sun's point of view, however, the lines would appear straight. Exactly how these lines bend depends on the shape and size of the ball. The same principle applies to AFI: An operator takes a picture of a subject covered in stripes of light (fringes) from a known angle. Knowing exactly where these fringes are projected from—the sun, in the venetian blind example—together with how they appear to bend as viewed from that angle, is enough to triangulate the height of each individual point.

3-D shoeprint In the near future, detectives may be able to compare 3-D photos of shoe soles (above) with corresponding 3-D images of shoeprints shot at crime scenes.

The system is still under development, but its potential applications are limited only by the imagination. (As Derr says, "If you can think of it, we can design it!") For crime-fighting, besides the possible uses already mentioned, Shirley and his colleagues have investigated using AFI to take images of suspects' ears, which in three dimensions serve as well as fingerprints in identifying individuals. Capturing a shoeprint in a patch of dirt would take less than a second; later, specialists could test whether 3-D measurements of a suspect's shoe matched the print.

Potential uses go far beyond detective work. Shirley and company fully realize how invaluable knowing the exact dimensions of a car door, satellite dish, or airplane wing, precise to a few hundred nanometers, would be to manufacturers. (Scalability is another of the system's strengths; the same principles that work on a square inch work just as well on an object measuring hundreds of feet square.) Imagine enhanced "virtual reality" games with photo-realistic, 3-D characters and places, or faxes and copying machines that offer printouts in three dimensions. With 3-D printers already on the market, such machines are anything but pie-in-the-sky.


Rob Meyer is Production Assistant of NOVA Online.

The technology reported here was developed at Lincoln Laboratory and was sponsored by the Department of the Air Force under Air Force Contract #F19628-95-C-0002. Opinions, interpretations, conclusions and recommendations are those of the author and are not necessarily endorsed by the United States Air Force.

Images courtesy of Lincoln Laboratory

Chronology of a Murder | Science in the Courtroom
Create a DNA Fingerprint | 3-D Mug Shot | Cleared by DNA
Resources | Transcript | Site Map

Editor's Picks | Previous Sites | Join Us/E-mail | TV/Web Schedule
About NOVA | Teachers | Site Map | Shop | Jobs | Search | To print
PBS Online | NOVA Online | WGBH

© | Updated November 2000

Support provided by

For new content
visit the redesigned
NOVA site