LEDs let archaeologists see the past in a new light

by Aarhus University

edited by Robert Egan

September 1, 2026

An archaeological excavation is a little like eating a layered cake: Once you have removed a layer, you cannot put it back. The problem becomes even trickier if two layers look almost exactly the same. Researchers from Aarhus University and Moesgaard Museum in Denmark have now developed a relatively inexpensive multispectral imaging system that gives archaeologists more ways of looking at the soil before they dig into it.

Instead of illuminating an excavation with ordinary white light, the prototype uses LEDs at 16 different wavelengths, ranging from near-ultraviolet through visible light to near-infrared.

Different materials absorb and reflect these wavelengths differently. Two deposits that both look like rather similar brown or black soil to the human eye may turn out to be quite different when viewed in another part of the electromagnetic spectrum.

"We wanted to make things faster, clearer and better in the field—and at a cost that makes sense for archaeology. One way of doing that is simply to rethink the light," says associate professor Søren Munch Kristiansen from the Department of Geoscience at Aarhus University.

The study is published in the Journal of Archaeological Science.

A difficult test at Sorte Muld

The researchers tested the first prototype at Sorte Muld on the Danish island of Bornholm.

Sorte Muld—literally "Black Soil"—is an archaeologically rich Iron Age site with more than a meter (more than 3 feet) of dark cultural deposits created by centuries of human activity.

For archaeology, that richness comes with a problem. Many of the layers are thin, irregular and very similar in color, making it difficult to determine exactly where one deposit ends and another begins.

The LED multispectral imaging system, or LEDMSI, photographed the same section repeatedly while illuminating it with the 16 different wavelengths.

The resulting images were then analyzed using statistical methods that combine the spectral information and enhance differences between materials.

Close-up photo of the prototype tested at Sorte Muld. The camera lens is in the center, surrounded by LEDs emitting light at 16 different wavelengths. Credit: Peter Gammelby, Aarhus University

And differences emerged that were difficult to distinguish in normal color photographs.

"The images looked almost psychedelic when we took them, and we only really saw what we had when we analyzed them back at the computer," Kristiansen says. "Then we could see new layers and changes in the fill that we simply could not see with the naked eye."

One of the features revealed by the analysis may represent an old ground surface and therefore a break in activity at the site. The researchers stress that the interpretation remains to be confirmed, but it illustrates what the system can do: point the archaeologist toward differences that might otherwise go unnoticed.

That matters because archaeological soil rarely resembles a neat pile of pancakes. People dig holes, fill them in again, move soil around and build on top of earlier activity. Two objects found in apparently separate layers may therefore be contemporary, while neighboring deposits may represent very different events.

Seeing those differences more clearly can improve both documentation and decisions about where to take samples for dating and other analyses.

This image is a so-called false color composite. The system scans the soil at different wavelengths, some of which are invisible to the human eye. A mathematical algorithm, ICA, then separates the combined signal and isolates distinct sources in the soil, displaying them in colors that we can see. The difference between the left and right images is that different values have been assigned to the red, green and blue channels. This makes it possible, for example, to see that the five yellow patches on the left are not all made of the same material: in the image on the right, the uppermost patch appears blue. Credit: David Stott, Moesgaard

Bones light up

The experiment revealed another useful trick.

When illuminated with ultraviolet light, small bone fragments fluoresced and became easier to distinguish from the surrounding soil.

Normally, fluorescence photography requires optical filters to separate fluorescence from reflected light. The researchers showed that useful information could also be extracted without such filtering, simplifying the equipment needed in the field.

This may eventually help archaeologists locate material worth sampling for scientific analyses. Previous research has linked fluorescence in archaeological bone to collagen preservation, which in turn can be relevant when selecting material for ancient DNA and protein analyses.

The present study does not demonstrate that LEDMSI can automatically select the best DNA samples. But it shows that fluorescent material can be detected quickly and with comparatively simple equipment.

The prototype is also considerably cheaper than hyperspectral imaging systems previously used for similar archaeological purposes—roughly 20 times cheaper, according to the researchers.

The new version of LEDMSI, with two LED flash units mounted on either side of a camera attached to the crossbar of the stand. The system requires complete darkness, so the whole setup is covered with black plastic when in use. Credit: David Stott, Moesgaard

From Bornholm to Viking Age graves

The system has already moved on from the version tested at Sorte Muld.

A second-generation prototype with a better camera and two LED units has recently been tested at Fredbjerg in northern Jutland, Denmark, where archaeologists are excavating a cemetery dating from the late Viking Age and early Christian period.

Preservation there is poor. In some graves, very little skeletal material remains, and the bodies are visible mainly as faint shadows in the soil.

"We're trying to see if there is information present in the graves that is not otherwise apparent," says archaeologist David Stott from Moesgaard Museum, one of the developers of the LEDMSI system.

During the Fredbjerg excavation, the researchers used the multispectral data to help decide where to take samples for further analyses.

The new version can acquire its series of images in around 30 seconds, compared with up to about five minutes for the first prototype. The researchers are now working toward processing the results onsite almost immediately on a computer—and eventually on a phone or other field device.

The LEDMSI prototype mounted in an excavation trench at Sorte Muld on Bornholm, shown here without the black plastic light shield. Credit: David Stott, Moesgaard

Next step: Machine learning

Another goal is to add machine-learning software that can learn the spectral signatures of different materials and help archaeologists work out what they are looking at.

"The difficult part is teaching an algorithm what is what. At the moment, the archaeologist and I, as a geologist, basically have two settings: glasses on or glasses off. This gives us many more ways of seeing," Kristiansen says.

He does not expect algorithms to replace the trained archaeological eye.

The ambition is to give that eye some assistance.

"I hope that in 10 years, bringing a system like this to an excavation will be as normal as bringing a digital camera," Kristiansen adds.