Photogrammetric 3D Digitisation of Bajuni and Pokomo Cultural Artefacts: A 65-Artefact Proof-of-Concept at Fort Jesus, Kenya

Source abstract

Digital ‌preservation ‌opens ‌a practical route for protecting cultural objects that face material decay, shifting environmental conditions, limited accessibility, and the gradual erosion of contextual knowledge. This article documents a practice-oriented deployment of photogrammetric 3D digitisation carried out under the Leveraging Artificial Intelligence for the Preservation and Promotion of Coastal Cultural Heritage in Kenya (LAIPCH-K) project. Attention is placed on artefacts accessed via the National Museums of Kenya at Fort Jesus in Mombasa, with particular emphasis on how photographic capture, technology choices, local computing capacity, 3D reconstruction, quality checks, and metadata handling were brought into a single production workflow. The production log covers 65 cultural artefacts, and every object is supported by photographic documentation. For 44 artefacts, explicit image totals were recorded, amounting to 7,283 photographs. Per-object counts span 80 to 210 images, while the average is about 166 photographs per artefact. Work proceeded through image sorting and inspection, alignment of camera poses, Structure-from-Motion processing, dense Multi-View Stereo reconstruction, mesh creation, texture generation, post-processing cleanup, model refinement for use and dissemination, and export. To determine an appropriate toolchain, three candidate options were assessed, Meshy multi-view image-to-3D, Agisoft Metashape Professional, and 3DF Zephyr Lite. The comparison considered image constraints, licensing and cost, where processing would occur, and the limits imposed by available computing hardware. On this basis, 3DF Zephyr Lite was chosen for the locally executed photogrammetry pipeline. Overall, the study frames museum digitisation under resource constraints as a socio-technical undertaking. Successful production depended not only on capture and reconstruction steps, but also on the coordination of software, computing infrastructure, technical know-how, metadata practices, and systematic quality assurance. The resulting digital assets form an evidence-linked base for later digital-heritage work, including metadata expansion, search and retrieval, classification, multilingual access, and immersive forms of engagement. While the work establishes operational viability and reports production outputs, it does not provide an independent assessment of geometric accuracy or metrological performance.