Edge detection of magnetic data: Preliminary results of application to Episkopeion region in Olympos Ancient City (Antalya)

Identifying and determining buried archaeological structure limits are crucial for archaeological prospection surveys. The archaeological prospection surveys can shed light on protecting cultural heritage. Magnetic is one of the most applied methods in archaeological surveys to plan the excavation process. It aims to identify buried temples, graves, city walls, and other structures by interpreting the data obtained from magnetic measurements, which is one of the non-destructive geophysical exploration methods. In the process of interpretation of magnetic data, many methods have been developed, such as Horizontal Gradient Magnitude, Analytical Signal, Theta Map, and Tilt Angle methods, which can generate information about the boundaries of potential subsurface features. By analyzing the gradients of the magnetic field data, inferences can be made to determine the edges of subsurface potential archeological structure distributions.

The Olympos is an important ancient city in Antalya (the south of Turkey) and contains many structures from the Byzantine period. This research aims to detect the pipe drain as a water system of the Episkopeion region in the Olympos ancient city. To this end, Tilt Angle (TA) and Edge detection field (ED) methods were tested on the magnetic map of the synthetic model. The horizontal boundaries of the potential pipe drain elements of the region were analyzed by applying the same procedures on the magnetic map obtained from the magnetic measurements on the area that has not been excavated in the Episkopeion archaeological excavation area. Magnetic measurements were carried out in 4 different regions within the Episkopeion area, and each area was numbered and analyzed separately. Combining the results obtained, an integrated visualization of the water system in the area was achieved.

Keywords: Archaeological Prospection, Edge Detection, Olympos, Tilt, Angle, Pipe drains.