A Technical Note on AI-Driven Archaeological Object Detection in Airborne LiDAR Derivative Data, with CNN as the Leading Technique

Archaeological research, driven by the quest to understand ancientcivilizations, relies heavily on detecting archaeological features to uncover hiddenhistorical information. This article explores the intersection of airborne (aerial) LiDARtechnology and Machine Learning (ML) techniques in archaeological feature detection.Airborne LiDAR, offering high-resolution 3D terrain maps, has significantly advancedarchaeological surveys by enabling the detection of ancient structures and landscapeswith improved accuracy and efficiency. ML algorithms, particularly ConvolutionalNeural Networks (CNNs), complement airborne LiDAR derivatives by automatingfeature detection and classification, thus enhancing the efficiency and accuracy ofarchaeological research. Through a comprehensive review of past studies, this technicalnote highlights the potential contributions of ML-based approaches on LiDARderivatives in detecting archaeological features, such as ancient settlements, burialmounds, and urban complexes. Despite notable advancements, challenges such as dataaccessibility, algorithm interpretability, and interdisciplinary collaboration persist.