From Visual Perception to Automated Detection: Integrating NDVI, SCP, and AI for the Identification of Archaeological Surface Markers

Filomena Papaleo, Simone Giosuè Madeo

Abstract

This paper presents a comparative analysis between traditional archaeological landscape interpretation and outputs derived from artificial intelligence and multispectral remote sensing. The study focuses on the middle Tanagro Valley, where NDVI (Normalized Difference Vegetation Index) and the Semi-Automatic Classification Plugin (SCP) for QGIS were applied to satellite and drone imagery to detect cropmarks and soilmarks indicative of buried structures. These results were compared with expert archaeological interpretations based on photointerpretation and historical data. The objective is to evaluate the predictive validity of automated processing versus manual interpretation and to assess the effectiveness of automated classification in reconstructing settlement patterns. Results highlight the potential of integrated approaches in enhancing archaeological detection, while also acknowledging risks of overfitting and false positives in complex landscapes.

Download

IMEKO-Metroarchaeo-2025-140.pdf

DOI

10.21014/tc26-2025.140

IMEKO TC

TC26 - Metrology for Cultural Heritage

Back to the proceedings