A low-cost imaging system utilizing 16 wavelengths of sunshine might assist archaeologists spot hidden layers and supplies earlier than excavation removes them.
Researchers from Aarhus College and Moesgaard Museum in Denmark developed the LED multispectral imaging system, or LEDMSI, to disclose variations missed by regular coloration pictures. The system makes use of LEDs starting from near-ultraviolet by way of seen mild to near-infrared.
Completely different supplies reply to every wavelength in several methods. In consequence, deposits with almost an identical colours underneath white mild might present clear variations in multispectral pictures.
The group examined the primary prototype at Sorte Muld on Bornholm, a wealthy Iron Age site with more than a meter of dark deposits built up through centuries of human activity. The site offered a hard test because many layers are thin, uneven and similar in color.
The system photographed the same soil profile under all 16 wavelengths. Researchers then used statistical methods to combine the spectral data and increase contrast between different deposits.
The processed images revealed boundaries and changes that were hard to see with the naked eye. One feature might mark an old ground surface and a break in activity, although the team says further work is needed to confirm the interpretation.
Such distinctions matter during excavation. Archaeological deposits rarely form neat layers. People dug pits, moved soil, filled holes and built over older activity. As a result, materials found in nearby deposits don’t always belong to the same event.
A clearer view of these boundaries could help archaeologists record layers and choose better places for samples. Those samples might later undergo dating or other scientific tests.
The system also showed promise for finding small pieces of bone. Under ultraviolet light, some bone fragments fluoresced and became easier to separate from the surrounding soil.
The researchers detected this fluorescence without the optical filters often used for fluorescence photography. This could simplify field equipment and speed up the search for material worth sampling.
Fluorescence in archaeological bone has been linked in earlier work to collagen preservation. Collagen is important when researchers select samples for ancient DNA and protein studies. The new work doesn’t show that LEDMSI itself identifies the best DNA samples, but it shows a fast way to locate fluorescent material.
Cost is another advantage. The researchers say their prototype is about 20 times cheaper than hyperspectral systems used for similar work.
The team has since tested a second version at Fredbjerg in northern Jutland. Archaeologists there are studying a late Viking Age and early Christian cemetery where skeletal remains are often poorly preserved.
In some graves, bodies survive mainly as faint traces in the soil. The multispectral images helped researchers decide where to collect samples for further study.
The newer system also works faster. The first prototype needed up to five minutes to capture its full image set. The second version does the job in about 30 seconds.
The researchers now want to process the images almost at once in the field. They also hope to bring the analysis to phones or other portable devices.
Machine learning is another planned step. Software could learn the spectral patterns of different materials and help archaeologists sort what appears in an excavation profile.
Søren Munch Kristiansen, an associate professor at Aarhus University, said the system gives archaeologists more ways to view soil. He stressed that algorithms aren’t meant to replace trained archaeologists.
David Stott, an archaeologist at Moesgaard, said the team is testing whether multispectral data reveal information in graves that ordinary images miss.
The researchers say more work is needed before LEDMSI becomes a broadly useful field tool. Improvements are still needed in the camera system, optical quality, data analysis and methods for classifying archaeological materials.
