Development of an image-based method for the detection of archaeological buried relics using multi-temporal satellite imagery

Abstract

Remote sensing has been successfully used for the exposure of shallow buried relics such as archaeological remains. The detection is mainly based on photointerpretation of high-resolution satellite or aerial images. Photointerpretation for archaeological purposes is focused on the identification of crop marks using visible and near infrared (VNIR) spectrum (e.g. vegetation indices) response, which is sensitive to vegetation stress. Detection of such marks is always performed through images of adequate spatial resolution, and therefore this procedure might be problematic in cases when there is a lack of accessibility to such kinds of data. This paper addresses this problem and illustrates an image-based method intended for the detection of crop marks using satellite data of inadequate spatial resolution. The overall methodology consists of seven separate steps. The method needs two areas of interest to be selected in the image, preferably in close proximity to one another. The first area is characterized as the ‘archaeological area under investigation’while the second is a vegetated non-archaeological area. These two areas are simultaneously examined in detail using spectral signatures, soil lines, and their phenological cycle characteristics. The proposed methodology has been successfully applied in three different areas in Cyprus and Greece, where the authors have already used the technique for validation purposes.