Recent advancements and the standardization of remote sensing techniques have led to a radical transformation in archaeological research. This progress has been made possible through the availability of advanced platforms, devices, and sensors, coupled with the increasing accessibility of open-access datasets. Remote sensing approaches, integrated with laboratory and field research, have become essential tools in archaeological investigations, enabling the study of ancient cultural dynamics. Concurrently, advancements in machine learning, particularly in data modeling and image classification, have expanded the analytical capabilities of academic research. These methodologies facilitate the rapid assessment of large datasets, enhancing the efficiency of pattern recognition and anomaly detection in the available data. This thesis explores the integration of common and advanced digital methodologies for identifying and classifying specific archaeological structures visible from orbital space. It also examines the relationships between these structures and the mobility, occupation, and subsistence strategies of the human groups that constructed them. The project’s first case study focuses on slab structures villages associated with hunter-gatherer groups from the Early and Mid-Holocene (7th- 5th millennia BCE) in the Egyptian Western Desert oases. The second case study investigates monumental constructions attributed to nomadic pastoralists in the Central and Western Sahara between the 5th-2nd millennia BCE. The third case study examines ditched villages and Neolithic communities on the islands of the central Mediterranean between the 6th-4th millennia BCE. Analyses were conducted using image enhancement, deep learning, and statistical approaches applied to panchromatic, optical, multispectral satellite imagery, and digital elevation models. The results of this study include georeferenced vectors of newly identified archaeological structures, updated spatial distribution maps of sites for the three case studies, and hypotheses regarding the mobility patterns of the human groups associated with each archaeological context. These findings demonstrate the effectiveness of non-invasive and computational approaches in supporting and complementing traditional archaeological studies.
La ricerca archeologica, in tempi recenti, ha subito una radicale traformazione dovuta al crescente avanzamento e standardizzazione delle tecniche di telerilevamento. Ciò è stato possible grazie alla disponibilità di piattaforme, dispositivi e sensori avanzati, unitamente all’aumento dei dataset open-access. Gli approcci di telerilevamento, integrati con le attività di ricerca laboratoriale e sul campo, sono diventati ad oggi fondamentali nelle indagini archeologiche, favorendo lo studio delle dinamiche culturali antiche. Parallelamente, i progressi nell’ambito del machine learning, specialmente nella modellazione di dati e nella classificazione delle immagini, hanno ampliato le possibilità analitiche della ricerca accademica. Queste metodologie permettono una valutazione rapida di grandi dataset, migliorando l’efficienza nell’individuazione di pattern e anomalie nei dati a disposizione. Questa tesi esplora l’integrazione di metodologie digitali comuni e avanzate per identificare e classificare specifiche strutture archeologiche visibili dallo spazio. L’analisi si concentra inoltre sull’esame delle relazioni tra queste strutture e la mobilità, l’occupazione e le strategie di sussistenza dei gruppi umani che le hanno costruite. Il primo caso studio del progetto concerne i villaggi di strutture a lastre associati ai cacciatori-raccoglitori del primo e medio Olocene (VII-V millennio BCE) nelle oasi del Deserto Occidentale Egiziano. Il secondo caso studio tratta le costruzioni monumentali delle culture pastorali nomadi del Sahara centrale ed occidentale durante il V-II millennio BCE. Il terzo caso studio si focalizza sui villaggi trincerati e le comunità neolitiche delle isole del Mediterraneo centrale durante il VI-IV millennio BCE. Le analisi sono state condotte con approcci di image enhancement, Deep Learning e metodi statistici applicati ad immagini satellitari pancromatiche, ottiche, multispettrali e modelli digitali di elevazione. I risultati di questo studio comprendono: vettori georeferenziati di nuove strutture archaeologiche, mappe con distribuzioni spaziali aggiornate dei siti dei tre casi studio e la presentazione di ipotesi sulla mobilità dei gruppi umani di pertinenza di ciascun contesto archeologico. Questi risultati dimostrano l’efficacia degli approcci non invasivi e computazionali a supporto e integrazione degli studi archeologici tradizionali.
Lo Sguardo della Macchina sulla Mobilità Preistorica: Applicazione del Machine Learning ai Dati Satellitari per l'Individuazione di Siti Archeologici nell'Italia Meridionale e nel Sahara
BRUCATO, ALESSIA
2025
Abstract
Recent advancements and the standardization of remote sensing techniques have led to a radical transformation in archaeological research. This progress has been made possible through the availability of advanced platforms, devices, and sensors, coupled with the increasing accessibility of open-access datasets. Remote sensing approaches, integrated with laboratory and field research, have become essential tools in archaeological investigations, enabling the study of ancient cultural dynamics. Concurrently, advancements in machine learning, particularly in data modeling and image classification, have expanded the analytical capabilities of academic research. These methodologies facilitate the rapid assessment of large datasets, enhancing the efficiency of pattern recognition and anomaly detection in the available data. This thesis explores the integration of common and advanced digital methodologies for identifying and classifying specific archaeological structures visible from orbital space. It also examines the relationships between these structures and the mobility, occupation, and subsistence strategies of the human groups that constructed them. The project’s first case study focuses on slab structures villages associated with hunter-gatherer groups from the Early and Mid-Holocene (7th- 5th millennia BCE) in the Egyptian Western Desert oases. The second case study investigates monumental constructions attributed to nomadic pastoralists in the Central and Western Sahara between the 5th-2nd millennia BCE. The third case study examines ditched villages and Neolithic communities on the islands of the central Mediterranean between the 6th-4th millennia BCE. Analyses were conducted using image enhancement, deep learning, and statistical approaches applied to panchromatic, optical, multispectral satellite imagery, and digital elevation models. The results of this study include georeferenced vectors of newly identified archaeological structures, updated spatial distribution maps of sites for the three case studies, and hypotheses regarding the mobility patterns of the human groups associated with each archaeological context. These findings demonstrate the effectiveness of non-invasive and computational approaches in supporting and complementing traditional archaeological studies.File | Dimensione | Formato | |
---|---|---|---|
AlessiaBrucato_37Cycle_PhDThesis_PDFA-signed.pdf
accesso aperto
Dimensione
120.93 MB
Formato
Adobe PDF
|
120.93 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/215061
URN:NBN:IT:UNIBA-215061