Detection of archaeological soil marks using airborne hyperspectral images

Soil marks are detectable by airborne images due to the difference in soil colour between the archaeological feature and the surrounding background soil. Colour of the soil only represents the visible part of the soil spectrum which contains physical and chemical information of the soil. This study will present a spectral analysis method to prospect soil mark features and buried archaeological remains using airborne image data. This method statistically calculates the difference between the targeted spectrum and the background (non-archaeological) soil spectrum. The difference is quantified by an R-value. If the R value is larger than 1, then the spectral behaviour of the targeted spectrum is different from the spectrum of the background soil and, thus, likely to be an archaeological soil spectrum (soil mark). In this study, the spectral analysis method will be applied to APEX imaging spectroscopy data collected from an archaeological site in Sárvíz Valley, Hungary. Previously, the method was successfully applied to the same archaeological site using soil spectra gathered by a portable hand-held VIS-NIR spectrometer. Here, the results showed clear spectral difference between soil mark features and background soil. This study will 1) compare the results of the method from hyperspectral image and ground-based spectral data, and 2) investigate the most effective waveband for identifying archaeological spectral signatures to verify the effectiveness of the method.