Aerial Remote Sensing for Precision Archaeology Using RGB–Multispectral Image Fusion of UAS Data

ABSTRACT: Precision Archaeology leverages advanced technologies, such as unmanned aircraft systems (UAS), for documenting archaeological sites with high spatial resolution and accuracy. This paper presents a reproducible RGB–multispectral (MS) image-fusion workflow for Precision Archaeology, combining PPK-based georeferencing with quantitative assessment of product accuracy and spectral preservation. Within this framework, the repeatability of the results produced by the UAS data fusion method confirms its reliability and establishes it as a valuable documentation tool. Among the experimental applications conducted to date, this paper adds two more: the Sanctuary of Eukleia at Aigai and the funerary ensemble in the Philippi plain, where Aerial Remote Sensing was performed using a UAS equipped with a Post-Processed Kinematic (PPK)–Global Navigation Satellite System (GNSS) receiver. A ground-based GNSS receiver was used to measure control points (CPs) and the base point used to correct the coordinates of the UAS image acquisition centers using the PPK method. For both archaeological sites, RGB and MS stereoscopic images were acquired from flight altitudes of 60 and 100 m, respectively, achieving an overall theoretical solution accuracy of under 2 cm. Digital surface models (DSMs) were generated with spatial resolutions of approximately 2 cm for the RGB and about 14 cm for the MS images, along with orthophotomosaics with spatial resolutions of roughly 1 cm for RGB and 7 cm for MS images. In the final stage, image fusion of the RGB and MS orthophotomosaics was applied, improving the spatial resolution of the MS orthophotomosaics from 7 cm to approximately 1 cm, while simultaneously preserving nearly all the original spectral information in the new fused images. Spectral preservation was quantified via band-wise correlation between the original MS and fused images (≈0.99 average for the Philippi dataset; ≈0.85 average for Aigai, likely influenced by a ~45 min RGB–MS acquisition gap and corresponding shadow/illumination differences). These new images can be used for classification purposes, enabling the identification of different materials and the detection of archaeological feature pathology with optimal spatial resolution and accuracy.