AI Driven Archaeological Site Mapping

Abstract— Identification of archaeological sites and their mapping is one of the key roles played in heritage preservation and supporting historical research. Most survey methods rely on ground exploration, which requires much time and resources. Still, all the same, are always limited by many accessibility concerns. This paper presents an AI-driven approach towards automatic archaeological site mapping from satellite images using a machine learning technique. Accordingly, a system is proposed that integrates all steps: image pre-processing, feature extraction, and classification based on deep learning to detect potential archaeological patterns such as structural remains, soil marks, and landscape anomalies. A convolutional neural network model will be trained on labeled geospatial datasets to enhance detection accuracy by reducing human intervention. Experimental results prove improved efficiency and more reliability compared to conventional survey methods, attaining a higher precision in identifying potential excavation zones. The proposed approach represents scalability, cost effectiveness, and no invasion in an archaeological exploration that enables faster decision making with enhanced preservation planning. Future work includes the integration of multi- spectral imagery and drone-based data acquisition, enhancing the detection performance and geographic adaptability.