Optimal Spatial Resolution for the Detection and Discrimination of Archaeological Proxies in Areas with Spectral Heterogeneity

Athos Agapiou

doi:10.3390/rs12010136

Abstract

Subsurface targets can be detected from space-borne sensors via archaeological proxies, known in the literature as cropmarks. A topic that has been limited in its investigation in the past is the identification of the optimal spatial resolution of satellite sensors, which can better support image extraction of archaeological proxies, especially in areas with spectral heterogeneity. In this study, we investigated the optimal spatial resolution (OSR) for two different cases studies. OSR refers to the pixel size in which the local variance, of a given area of interest (e.g., archaeological proxy), is minimized, without losing key details necessary for adequate interpretation of the cropmarks. The first case study comprises of a simulated spectral dataset that aims to model a shallow buried archaeological target cultivated on top with barley crops, while the second case study considers an existing site in Cyprus, namely the archaeological site of “Nea Paphos”. The overall methodology adopted in the study is composed of five steps: firstly, we defined the area of interest (Step 1), then we selected the local mean-variance value as the optimization criterion of the OSR (Step 2), while in the next step (Step 3), we spatially aggregated (upscale) the initial spectral datasets for both case studies. In our investigation, the spectral range was limited to the visible and near-infrared part of the spectrum. Based on these findings, we determined the OSR (Step 4), and finally, we verified the results (Step 5). The OSR was estimated for each spectral band, namely the blue, green, red, and near-infrared bands, while the study was expanded to also include vegetation indices, such as the Simple Ratio (SR), the Atmospheric Resistance Vegetation Index (ARVI), and the Normalized Difference Vegetation Index (NDVI). The outcomes indicated that the OSR could minimize the local spectral variance, thus minimizing the spectral noise, and, consequently, better support image processing for the extraction of archaeological proxies in areas with high spectral heterogeneity.

Highlights

Since the launch of the first commercial high-spatial-resolution multispectral satellite sensor, namely the IKONOS, archaeological prospection surveys systematically explored spaceborne optical-based observations [1,2]
A topic that has been limited in its investigation in the past is the identification of the optimal spatial resolution of satellite sensors, which can better support image extraction of archaeological proxies, especially in areas with spectral heterogeneity
The optimal spatial resolution (OSR) was estimated for each spectral band, namely the blue, green, red, and near-infrared bands, while the study was expanded to include vegetation indices, such as the Simple Ratio (SR), the Atmospheric Resistance Vegetation Index (ARVI), and the Normalized Difference Vegetation Index (NDVI)

Summary

Introduction

Since the launch of the first commercial high-spatial-resolution multispectral satellite sensor, namely the IKONOS, archaeological prospection surveys systematically explored spaceborne optical-based observations [1,2]. To better understand an archaeological landscape, a particular focus that has attracted the interest of the scientific community was the exploitation of such space-based observations for the detection of shallow buried remains. The latest can be achieved through the detection of archaeological proxies, known in the literature as cropmarks [6,7,8]. Such archaeological proxies (cropmarks) detected in satellite images can be used to map and detect subsurface archaeological remains at shallow depths. Cropmarks are usually formed in areas where vegetation overlay shallow-buried archaeological remains. As [15] stated, depending on the type of the buried archaeological features, crop vigor may be enhanced or reduced

Objectives

Methods

Results

Conclusion