Work

Lightweight Metadata For Leveraging Crowdsourced 3D-Representations

Year: 2026
Type: article
Abstract: The proliferation of digital artefacts in cultural heritage presents both unprecedented opportunities and poses additional challenges. Diverse data formats, ranging from textual descriptions and photographic documentation to architectural drawings and 3D models, are being produced by a wide variety of professionals involved in safeguarding heritage. Even within 3D representations, significant variations exist in quality, level of detail, and semantic depth: models may range from partial, semantically limited point clouds to richly annotated Industry Foundation Classes (IFC) files. In addition, several trends are reshaping the domain: Linked Data enables machine-readable representation of complex relationships between events, actors, sites, and their digital counterparts; Historic Building Information Modelling (HBIM) is maturing into a conservation methodology (Penjor et al., 2024); widely available and constantly improving consumer technologies facilitate high-quality data acquisition, including 3D reality capture. Moreover, in heritage preservation, a shift in paradigm is taking place, characterised by the growing prominence of community-oriented approaches (Campfens et al., 2023, p. 17). This research is motivated by an urgent need to systematically document Ukrainian cultural heritage in light of the ongoing Russo-Ukrainian war. Preparatory measures, such as inventories and the digitisation of notable assets, were not fully implemented before the war began (Campfens et al., 2023, pp. 65-66). As access to heritage sites and archives containing their paper-based documentation is now partially restricted due to military operations, the inaccessibility of fundamental architectural documentation hinders restoration efforts. In the Ukrainian context, crowdsourcing represents a promising strategy for large-scale (remote) data collection, as it broadens public engagement with local heritage while helping mitigate the shortage of specialists and financial resources that frequently accompany armed conflicts. However, the resulting multiplicity and variety of contributors and data require approaches that ensure traceability of evidence and compliance with FAIR (Findable, Accessible, Interoperable, Reusable) principles (Wilkinson et al., 2016). In large-scale and time-sensitive contexts, scalable solutions are needed for both distributed data acquisition as well as their efficient search, retrieval, and reuse. This calls for the adoption of machine-readable structures and open formats for coordinated documentation. Rich metadata allows professionals to pre-evaluate the suitability of 3D representations without inspecting each file individually and select data tailored to specific analytical needs from a vast pool of documents, thus reducing manual effort. For example, architects may look for detailed as-planned models with extensive information about structure´s materials to support restoration planning, while a law enforcer would rather be interested in gathering evidence of Russian crimes in the form of geometrically precise point clouds with timestamps showcasing the results of recent bombing; an urban analyst might instead query for multiple low-resolution, mobile scans within a specific neighbourhood to reconstruct a broader context. All in all, sensible and coherent metadata, if recorded consistently, enables managing digital artefacts at scale, catering to diverse use cases within heritage safeguarding activities. This research creates a minimal yet expressive set of metadata for 3D representations of immovable cultural heritage, designed to describe the essential reality capture concepts, such as acquisition and post-processing methods, model scope and resolution, provenance information, etc., while remaining lightweight to be populated by laypeople. Thus, it bridges the gap between high-quality data requirements and low-threshold participative collection.
Language: English
Subfield: Geology
Open Access status: green
OpenAlex ID: W7225101786
FWCI:
Cites:
Cited by:
Related to: 10