Archaeological Surveys With Drone Mapping

An archaeology researcher in the United Kingdom uses drones and professional multispectral sensors to detect buried archaeological features previously seen only after time-consuming geophysical ground surveys.

Using multispectral imaging to detect ancient man-made structures

Agricultural fields

Agriculture and archaeology: two disciplines that typically are not thought to go together. But in regions that are rich in archaeological features, the link can be strong.

Soils defines how plants grow and the characteristics of soils are many times affected by man-made features that are buried below the ground. This interaction leads to what are called “cropmarks”. Sometimes these effects are noticeable to the naked eye. Many times, the soil variability is subtle and results only in small changes in the vigor of the plant or the chlorophyll content in the leaves. That small variability (if natural or man-made) can be detected through remote sensing by using multispectral sensors, especially advanced sensors that are able to detect changes in chlorophyll content in leaves.

Features buried under ground
Archaeological features buried under ground can affect plants that grow above them. Plants that grow on soil above the remains of a foundation for a wall (left) will generally not fare as well as plants that are above the remains of an old ditch that has been filled in (right).

How multispectral imagery fits in

For the past year and a half, Henry Webber, a Ph.D. candidate at the University of Bristol’s Department of Archaeology and Anthropology, has been gathering data using a drone that carries a MicaSense RedEdge multispectral sensor. He has flown over agricultural fields that have known archaeological features beneath them and has compared his findings with more conventional tools for surveying such as a magnetic gradiometers, which detect changes beneath the surface by measuring the strength and orientation of the magnetic field of the ground below.

Henry Webber holds the drone and the RedEdge multispectral sensor
Henry Webber holds the drone and the RedEdge multispectral sensor he used in his research. Photo courtesy Henry Webber.

A winter wheat field near the village of Wilsford in the English county of Wiltshire provides an example. The field contains a Neolithic henge (a Stone Age monument that is thought to have been used for rituals and burials) as well as a Roman era farmstead. Henry has been mapping archaeological features in this study area as part of his Ph.D. research using a variety of geophysical surveying tools. The magnetic gradiometer surveys have yielded impressive results, revealing in high detail the structure of archaeological features beneath the surface. But the information acquired with multispectral remote sensing has not only largely replicated these surveys but has also shown a previously unknown detail.

Referencing the figure below, the image on the left, obtained with the RedEdge sensor, is a representation of a vegetation index in MicaSense Atlas called the Chlorophyll Map. This index uses multiple spectral bands, including the Red Edge and Near-Infrared bands (not visible to the human eye) and is generally correlated to chlorophyll content on leaves. Brighter areas in the map indicate areas where the winter wheat has slightly higher index values (and correspondingly higher levels of chlorophyll). The image on the right comes from the magnetic gradiometer survey and shows an L-shaped structure under the ground.

A rectangular enclosure buried underground becomes visible when using multispectral data and vegetation indices sensitive to chlorophyll content (left), revealing features that were not directly visible using conventional geophysical survey tools (right).

RGB color image of the Wilsford enclosure for reference.

The multispectral imagery shows that the boundaries of this archaeological feature extend to the bottom and to the right, forming what appears to be a rectangular enclosure. This under-ground feature, not directly visible in the geophysical survey, becomes evident when the effects of changing soil affect the health of the plants above. “This discovery invites a number of exciting questions,” says Henry. “Is this enclosure part of the Neolithic henge? Or is it part of the Roman farmstead, making it much larger than originally thought?” Henry is hoping that future archaeological surveys in this area will help answer these questions.

In the same wheat field, a large neolithic henge is clearly visible. A number of small circular pits lie buried just below this feature. While inspecting the results from the multispectral data, Henry noticed that two of the three pits were clearly visible in the chlorophyll map output but not in the geophysical survey. Henry believes these findings can help guide researchers to understand the circumstances under which magnetic-sensor based surveys sometimes miss certain archaeological features.

Another example of multispectral data complementing geophysical surveys. Here, two circular pits below the henge are visible in the chlorophyll map image (left) but not in the map generated with magnetic gradiometer surveys (right).

The ability to see new features and potential anomalies and to compare existing surveys with new ones opens new possibilities for understanding more about the ancient structures that lie beneath our feet–Henry Webber, Archaeology Researcher

A Roman villa emerges through a wheat crop

Near the village of Kingscote in Gloucestershire, England (about 100 miles east of London) lie the buried remains of a Roman settlement. Dating back to between the 1st and 4th century AD, this archaeological site is comprised of approximately 75 buildings and structures. One of the most intriguing structures in this site is a Roman villa that lies underneath a wheat field.

RGB color image of the buried villa for reference.

Henry captured multispectral data over this site and was surprised to see the results. The figure below (left) shows an image acquired with the multispectral camera, featuring the chlorophyll-sensitive vegetation index NDRE. An overlay of the outline of the Roman villa (obtained from archeological surveys and excavation) is shown on the right. For comparison, the standard RGB color image of the same area is shown as well. “It is exciting to see how well the NDRE images match the geological survey results,” says Henry. “Features that are invisible or very hard to see with the naked eye or in color pictures show up clearly in the multispectral data”.

A promising set of first results, and more to come

Henry is looking forward to continuing to explore the capabilities of multispectral imaging using advanced sensors as a tool to complement other archeological tools. “It is clear that there is value in this type of imagery, even in areas that have been mapped before,” Henry explains. “The ability to see new features and potential anomalies and to compare existing surveys with new ones opens new possibilities for understanding more about the ancient structures that lie beneath our feet.”

Search
Generic filters
Filter by Category
Case Studies
How It Works
How To
News

POPULAR TAGS:

VIEW BY POST TYPE:

SEARCH BY DATE:

Archives

RECENT POSTS:

FEATURED VIDEO: