Spaceborne Remote Sensing for Near Eastern Archaeology: A case study on archaeological site-detection in Jordan's Black Desert

The Black Desert in the east of Jordan has recently gained attention in the archaeological domain due to the discovery of pre-historical structures and a vast amount of rockart depictions. The Department of Near Eastern Archaeology of the University of Leiden has started investigating the cultural developments within this area from c. 1000 BC to 500 AD under the ‘Jebel Qurma Archaeological Landscape Project’. The main archaeological structures of interest are known as ‘The Works of the Old Men’. These structures are clearly visible from an aerial perspective, but hard to recognize at ground level. Furthermore, they are extensive in scale as they are known to occur throughout the whole Arabian peninsula. Until now, structures have been mainly documented based on human visual inspection of aerial and spaceborne optical remote sensing. This research covers the results of a case study on the use of spaceborne optical and radar remote sensing imagery for archaeological site detection in the Near East in an automatic way. This will facilitate faster analysis of the area of interest, as well as the Arabian peninsula as a whole. For the research, imagery from several spaceborne optical and radar imagery was acquired, most notably from the CORONA, IKONOS, ALOS and TerraSAR-X satellites. Fieldwork was performed in the region of interest as well, to place corner reflectors and perform GPS measurements of structures of interest. As the corner reflectors were captured in the last two acquisitions of the ordered TerraSAR-X imagery, they served to improve the geo-referencing of the radar imagery. All imagery was processed to form a co-registered, geo-referenced imagery stack for use in feature detection. Based on observations during fieldwork and the literature available, the main structures of interest were parametrized. An automated test using the GPS measurements was used to determine the suitability of layers of the imagery stack for automatic feature detection based on their normalized cross correlation. However, the currently obtained radar imagery proved to have insufficient spatial resolution to allow for automatic feature detection. Two different kinds of detection algorithms were designed for feature detection: the Direction of Constant Gradient (DoCG) and the Laplacian of Gaussian (LoG) approach. Of these two, the LoG approach performed better in the detection of structures. The end result provides a faster way to analyze large areas in an intuitive way. Possible archaeological sites can thus be directly identified and documented.

Subject

archaeology
site-detection
remote sensing
SAR
GPS
corner reflectors
geo-referencing
image processing

To reference this document use:

http://resolver.tudelft.nl/uuid:e863c6c6-852e-42d1-9053-9a014727e74b

Part of collection

Student theses

Document type

master thesis

Rights

Files

PDF

Spaceborne_Remote_Sensing ... _Liem_.pdf

38.11 MB

Close viewer