EEs dig up imaging technology to aid archaeologists

By Adam Marcus
EE Times
(07/09/01, 11:13 a.m. EST)

	Linux desktop push could benefit disabled   Sharp demos 5.6-MHz 1-bit audio amp   Altera kicks off mask-programmable PLD program   Merger could streamline India's tangled telecom sector   Design vendor Forte cuts staff

Visitors to the Beckman Institute on the campus of the University of Illinois in Urbana-Champaign might stumble onto a bizarre ritual: In a laboratory, a group of earnest engineers standing in a sandbox shoots bursts of sound waves toward the floor. Filling the room are audible, and somewhat nerve-wracking clicks, which explains the airport ear guards worn by the researchers.

They're not trying to stun earthworms. Rather, the researchers are acting out a marriage of sorts between Indiana Jones and Heinrich Hertz, a union that enlists acoustical imaging technologies in the search for the ancient.

Leading the effort is William O'Brien, a professor of electrical and computer engineering. O'Brien, whose work in the past has focused on acoustic imaging for medical devices, was approached in the mid-1990s by the Pentagon, which wanted to see if the technology could be adapted to its archaeological endeavors. Like other government agencies, the military is required to conduct archaeological surveys of the land it hopes to develop or train on to make sure it's clear of culturally significant artifacts.

Conventional metal detectors are a first step, but not a particularly thorough one. "They miss fair amounts with that technique, because most archaeological objects tend to be non-metallic," O'Brien said.

Ground-penetrating radar has also been a big help in this area, but its utility is limited to dry ground. "The ground-penetrating radar can go pretty deep if the soil is dry, but even a day after it has rained there's enough moisture to prevent the electromagnetic energy from going too deep to reveal much," O'Brien said.

While acoustical imaging could in theory overcome the problem of water — sound waves love the liquid — it wasn't at all clear that the technology would be worth much in soil. "We didn't know how lossy the soils would be. If they're very lossy, when distance and the attenuation is too great then the signals wouldn't go far," O'Brien said.

The Illinois team chose to work with sand, which is a "harsh" material from the perspective of signal loss but has the twin virtues of not layering like soil and being easy to bury objects in, O'Brien said. So they built a sandbox, five cubic feet, in the bioacoustics research lab and began hiding things, from steel pipes to baseballs, like a pack of two-footed squirrels.

A 5-kHz signal allows the researchers to image objects up to a foot beneath the surface, while 1 kHz can look twice as deep. At those depths, they are comfortably resolving objects about 4 centimeters in diameter. The goal is 1 cm.

Stick figure

"We're looking for pretty fine resolution," said David C. Munson Jr., a professor of electrical and computer engineering at Illinois who helped design the image processing side of the tool, called a transducer. An object "could be a Roman coin. We're not proposing to make an image of the coin, just that there's something there. Maybe it's a skeleton of a person. We'd at least like enough resolution to see that it's a skeleton and not just some sticks."

On the other hand, Munson said, "Ultimately we'd like to produce an image that had at least some visual similarity to an optical photograph. So far the resolution is not quite what we would be hoping for."

"There's a typical engineering trade-off in almost all acoustical imaging," said O'Brien. "As frequency increases, the attenuation coefficient increases. It's exactly the same principle for why most medical ultrasound devices work at 1 to 10 MHz. As you go higher in frequency you can't penetrate as deep into the material." For now, he said, 1 cm is possible, but only if the team is willing to stick to within a foot of the surface.

Of course, many significant archaeological finds lie well below the early grades of topsoil. "But they tell me that they can learn a lot about whether or not to dig deeper based on what's just underneath the surface," O'Brien said.

The group has been considering linear arrays of transducers, perhaps three or four feet long yet quite narrow in the second dimension. The slim side would sacrifice resolution, but that could be overcome by moving the whole contraption and by transmitting and receiving in multiple directions, Munson said. "And by signal processing we're able to synthesize a larger area."

O'Brien and his colleagues are looking at ways to mount the transducer, which weighs about 100 pounds, on a mobile unit. "It would not be portable in the sense of a human carrying it around, but it could be like a pushable lawnmower," he said.

Charles Marsh, a materials scientist with the Army Corps of Engineers' Construction Engineering Research Laboratory in Urbana-Champaign, commissioned the project in 1992. Spotty funding has waylaid the work over the years, but Marsh is pleased that the work has come this far. The Corps hopes to test the transducer in November at its controlled archaeological test site, a simulated cultural dig with buried objects that have been fully mapped.

Then there are the archaeologists, whose field stands to gain greatly from the increasing sophistication of imaging technologies.

Avoiding destruction

Patrick Ryan Williams, a professor of archaeology and remote sensing at Boston University, said the latter pursuit can help the former resolve an inherent irony. Archaeologists "are focused on trying to get as much information from areas as we can, but traditionally the process that we have had to undertake is a destructive process. A lot of the new technology since the 1970s has been really enhancing our ability to learn about what is beneath the ground without having to destroy it."

Even if an artifact isn't damaged in the act of digging, its relationship to its setting is often disturbed, Williams said. "Remote sensing provides us a really rapid way of assessing an area without much expense."

Unfortunately, said Williams, the current subsurface imaging technologies provide resolution that's simply too crude to be of much use deep down. "One of the major things that could come out of this [acoustic] work is deeper sounding but maintaining that fine-scale resolution. That is going to be the very exciting thing for us."

New archaeologists are being trained in imaging methods and some universities are making it a priority to hire faculty experts in electronics. But Williams estimates that while as many as half of today's field archaeologists are interested in the technologies, only about one in 10 regularly uses them in their surveys.

Kenneth Kvamme, associate professor of anthropology at the University of Arkansas, is among the vanguard in adapting imaging techniques to his discipline. Kvamme, director of the Fayetteville school's ArcheoImaging Laboratory, is a specialist in Native American settlements in the Great Plains and he's one of the few to endorse imaging.

"The geophysical applications are in their infancy in North America," Kvamme said. They're also expensive, very much so by humanities department standards, since a radar system can cost $30,000, or the salary of a graduate student, which makes them even scarcer on campus.

Lately, Kvamme has been using his imaging tools in North Dakota's sprawling Whistling Elk site. From the surface, there's no hint of the once-active settlement below, he said. But when he and his research team canvassed the area for electro-resistivity, they detected some 60 ancient houses, a large fort complete with bastions, and a large squarish structure of obvious importance to the town.

"Electro-resistivity found that house really well," he said. "Inside the house it was filled in with sand, so it was more resistant and stood out really well. The geography was so accurate I put a stake in the ground and said that's where the hearth is and another where the entry way was, and we were within centimeters of it."

Sponsor Links National Semiconductor & iAppliance WebCast Click Here For A NetSeminar on New Architecture for Internet Devices! Synopsys, Inc. Presents a Free Net Seminar- Reducing the SoC Design Cycle with SystemC-based CoCentric® Tools Numerical Technologies The subwavelength barrier is gone, and 0.10 micron silicon is just the start. Microsoft Attend the Microsoft Windows Embedded DevCon, November 27 - 30, 2001