Algorithms for Enhancing Satellite Imagery to Discover Archaeological Finds Covered by Shadow

Abstract

Very high-resolution (VHR) images proved to be an invaluable source of information even in the archaeological domain, but sometimes shadows hinder their full exploitation. To overcome such limitation, this research proposes a workflow able to analyze shadowed zones, by processing Pléiades and World-View 2 images. The case study is the archaeological site of Maltai, in the Iraqi Kurdistan Region, which presents shadowed areas to be detected. Applying de-shadowing workflow has been tested over multispectral and panchromatic images, with different invariant color spaces. The proposed methods exploit the techniques of automatic thresholding and spectral ratio in the detection of shadow regions. This approach shows a clustering of shadow pixels for an enhanced images visualization and proves its suitability for archaeological settings.