Utilizing Low-Altitude Imagery and Weak Learner Algorithms for the Detection of Archaeological Surface Ceramics: The Kophinou case study

Abstract

In recent years, the application of remote sensing in archaeological research has grown significantly [1,2]. This study builds upon a 2021 case study conducted in Kophinou village, Cyprus, with a focus on improving the detection of archaeological surface ceramics using low-altitude imagery. The initial approach successfully identified scattered archaeological objects, even when they exhibited high spectral similarity to the surrounding surface [3]. This study further explores the potential of cost-effective, low-altitude sensor data and 2 machine learning (ML) algorithms to accurately detect and map archaeological surface ceramics. However, challenges were encountered in accurately detecting ceramics, particularly due to the issue of imbalanced data. To address this problem, a new methodology was developed to enhance the detection of surface ceramics. The aim of this research is to address this gap in the literature by incorporating AI methodologies to manage non-uniformly distributed classes, using weak learner algorithms to provide a more efficient approach in terms of both time and accuracy. The methodology presented here has practical applications for improving archaeological foot surveys.