Pretty Pictures and Research Tool: Enhancing 3D Documentation of Historic Clinker Boats Through Parametric Modeling

Pretty Pictures and Research Tool: Enhancing 3D Documentation of Historic Clinker Boats Through Parametric Modeling Fröhlich Fred1, Nielsen Sigurd Solhaug2 1 Department of Game Development, University of Inland Norway – Hamar, Norway 2 Glomdalsmuseet, Anno museum – Elverum, Norway ABSTRACT Recent advances in photogrammetry and LiDAR have transformed the digital preservation of archaeological artifacts by capturing detailed surface geometries. Yet, such methods often overlook the underlying construction principles that define complex objects, such as historic clinker-built boats. With this paper we present a case study from the Mjøsbåt Project, a conservation initiative for early 20th-century clinker-built boats in Norway, demonstrating how parametric modeling in SideFX Houdini can complement photogrammetric documentation. By overlaying scans with procedural reconstructions, we capture not only the visual form but also the logic of clinker boat construction, including plank overlap, bending, and alignment. The resulting models enable the generation of 2D drawings, physical scale models, and interactive assets while preserving intangible craft knowledge recognized in UNESCO’s listing of Nordic clinker boat traditions. We argue that aesthetic quality and research utility are not mutually exclusive but mutually reinforcing, and propose that parametric reverse engineering offers a pathway toward more comprehensive digital heritage practices. Beyond clinker boats, this approach has potential applications for documenting other complex archaeological artifacts where construction processes are as important as surface detail. Keywords: 3D Visualization, Reverse Engineering, Procedural Modeling, Clinker Boats, Photogrammetry, Digital Heritage, Virtual Archaeology Introduction Photogrammetry and LiDAR have become standard tools for recording archaeological artifacts and cultural heritage objects, offering highly detailed surface geometries that benefit both research and public engagement (Lanjouw, 2016; Pritchard et al., 2021). Yet these methods primarily replicate appearances and often fail to capture construction principles and material logics that are central to understanding complex objects (Chalmers, 2002). This limitation is especially evident in historic clinker-built boats, where overlapping planks, joinery, and shaping techniques embody craft traditions as much as material form. As Leijonhufvud (2022) notes, both manual and digital boat documentation methods have often overlooked the craftsperson’s perspective, reflecting instead that of the archaeologist or ethnologist. He argues that documentation should be informed by reconstruction and by an understanding of how a boat was made and assembled—not merely by its outward form. The boatbuilder’s viewpoint, shaped by embodied experience and tacit knowledge, provides an essential interpretive framework for reading both physical artifacts and digital models. In this study, we seek to integrate that perspective through computational means. By encoding construction logic within parametric models, we explore how digital tools can preserve and reactivate craft knowledge as an integral component of 3D documentation. Surface-based methods also face technical challenges. Photogrammetry and LiDAR are prone to issues such as calibration errors, occlusion, inaccessible spaces, and lighting difficulties, which can produce incomplete point clouds with holes and distortions (Varady et al., 1997). For example, scanning beneath the seats of the boats often yields missing data that prevents accurate reconstructions. These shortcomings highlight the need for approaches that move beyond surface fidelity to also document underlying construction logic. Figure 1 - Screenshot with mesh inconsistencies in photogrammetry data of a digitally reconstructed boat. To address this gap, we apply reverse engineering as both a conceptual and technical framework. In our workflow, photogrammetry provides a baseline record, while construction principles are distilled through expert knowledge and formalized into algorithms within a Houdini Digital Asset (HDA). This hybrid approach makes it possible to generate parametric 3D models that not only reproduce form but also encode the rules of construction. The aim of this study is to uncover the underlying construction principles of historic clinker boats and to produce clean, versatile 3D reconstructions. These models can be used for conservation, 2D plotting, printing scale models, structural analysis, and interactive applications. By bridging aesthetics and research, this work aligns with broader goals of digital heritage, including UNESCO’s recognition of intangible traditions in Nordic clinker boatbuilding. This study tackles the challenge of extending photogrammetric documentation with parametric modeling in order to preserve not only the external appearance but also the embedded construction knowledge of clinker-built boats. Background & Related Work Photogrammetry is widely used in archaeology for documenting artifacts and complex objects, valued for its low cost and accessibility. Protocols such as SOAP and HRP (Cerasoni et al., 2022) enable the creation of detailed, visually rich models that are especially effective for education and dissemination. Yet the technique is highly sensitive to lighting, surface properties, and occlusion, which can leave gaps or distortions in the data. Originally developed for landscapes and architecture, LiDAR is now applied at object level with portable scanners. It produces dense point clouds independent of lighting, making it useful for shaded or obstructed features such as boat interiors. Advantages include geometric accuracy and speed, though drawbacks are high equipment cost, specialist expertise, lack of native color, and difficulty with very fine or reflective surfaces. Photogrammetry and LiDAR are increasingly combined to leverage their strengths: photogrammetry provides texture realism, while LiDAR ensures geometric accuracy. This hybrid workflow is particularly effective for documenting complex heritage objects, where both appearance and structure must be preserved. Despite these advances, surface-based methods remain limited. As Jeffrey (2015) notes, heritage objects carry not only physical form but also histories, sensations, and craft knowledge that cannot be reduced to surface appearance alone. Digital models that replicate only visible geometry risk becoming “sanitised entities,” disconnected from the intangible and experiential qualities of heritage. Another challenge is the lack of standardization across digital heritage projects. As Pritchard et al. (2021) observe, 3D models are often complex and inconsistent in quality, making integration difficult. Geometry cannot be treated in isolation; it must be contextualized with materials, construction techniques, and multiscale information in order to support research and conservation. 3D modeling has been explored as a way to move beyond static surface scans. Tools like Autodesk Maya and Rhino/Orca 3D have been used for naval reconstructions (Yamafune et al., 2016; Tanner, 2013, Tanner, 2017; McCarthy & van Duivenvoorde, 2021). While these approaches allow some automation, they are limited in flexibility when repetitive elements such as planks or nails require modification. More recent studies have employed procedural modeling in SideFX Houdini. Suarez et al. (Suarez et al., 2019) demonstrated its versatility in reconstructing the lower hull timbers of a sixteenth-century merchant ship, noting that such procedural workflows supported experimentation and interpretation in ways unthinkable with conventional software. However, even in this promising work, challenges remain. Suarez et al. (2019) concluded that while their workflow was successful, the planking—arguably a very challenging aspect of the shipbuilding process—remained unresolved. This gap is significant: clinker-built boats depend on overlapping planks and subtle joinery that encode craft knowledge central to their construction. Current documentation methods therefore struggle to integrate both visual fidelity and construction logic, underscoring the need for new approaches. Building on this trajectory, our study develops parametric reconstruction methods that directly address the planking problem. By encoding boatbuilding rules into procedural models, we extend digital documentation from replicating surface appearance toward capturing the construction principles that define clinker boat traditions. These challenges set the stage for our work with the Mjøsbåt Project, where documenting the clinker boats of Lake Mjøsa offers a unique opportunity to test new parametric methods while also engaging with a living heritage tradition. Case Study Context: The Mjøsbåt Project Anno Museum is a consortium of small and medium-sized museums in Hedmark, Inland County, Norway. The Mjøsbåt Project grew out of the museum’s ambition to build competence in 3D documentation as a way to improve heritage management both within and beyond its collections. While the museum holds several traditional clinker boats, much of the tangible and intangible knowledge surrounding their use and construction remains embedded in the local community, where quite a few boats still serve cultural and practical purposes. The project therefore combines 3D documentation, experimental boatbuilding, and the continued use of local vessels; in this paper, we focus on the first two aspects. A key experiment involved documenting a traditional boat from 1946 and using its 3D model as the basis for constructing a replica—without the boatbuilder ever handling the original. This simulated a scenario in which an artifact is lost but its digital proxy survives, challenging both museum staff and the boatbuilder to reconstruct heritage collaboratively from digital data. The goal was not to reproduce dents and imperfections but to achieve a functional workboat, nearly identical in size and shape to the original, suitable for use by the local population. Figure 2 - Publicly accessible workshop on replica fabrication, summer 2024. The boats in question are clinker-built fishing and workboats associated with Lake Mjøsa in the interwar period. Typically 4.5–6 meters long and 1.25–1.40 meters wide, they are lightly built in spruce with 5–6 planks on each side of a two-part T-shaped keel. The planks overlap by roughly 20 mm and are fastened with galvanized nails and rivets; in some cases, simply bent nails suffice. Planks are bevelled to fit flush with the stem, which is often quarter-circle shaped and constructed in two parts. Frames vary from heavier sawn timbers to lighter combinations of sawn and steam-bent oak. Unlike other Norwegian inshore craft, these boats rely less on naturally curved wood, suggesting pragmatic choices driven by economy and reducing building time (Færøyvik et al., 1979). Construction methods also varied, with some built over molds upside down, but most with the keel set on the floor. The type appears to have been standardized in the 1920s, and coincided when leisure time and recreational pursuits expanded (Kjeldstadli, 1993; Olstad, 2022). Many surviving examples feature distinctive wrought-iron outriggers, which extend the oarlocks about 20 cm from the hull—a design element not found elsewhere in Norway but likely inspired by international racing boats introduced to Lake Mjøsa in the late 19th century (Hoem, 2003). Figure 3 - Mjøsbåt near Jessnes, ca 1935-1937. Photo: Johs Johannessen / Anno museum. Inventory no.: 0401-10204 During the interwar years, rowing regattas became popular social events, often organized by fishing associations and later by dedicated rowing clubs. Reports document distances, times, and even the names of competing boatbuilders, underscoring the close connection between craft, sport, and community identity (Nielsen, in press). Although the heyday of these regattas was brief, their cultural imprint is significant and shapes how we approach the preservation of these boats today. As Jeffrey (2015) cautions, digital heritage must remain grounded in historical context, ensuring that technologies serve both tangible preservation and cultural memory. The Mjøsbåt Project aims to document around 30 clinker-built boats using photogrammetry, most of which are privately owned and distributed across the district. Supported by the Directorate for Cultural Heritage, the project builds on existing local expertise and earlier analogue fieldwork in the region. Between 1997 and 2000, the Norwegian Forest Museum documented more than 450 workboats of various types throughout the Inland region (Skjærbakken, 2005). Extending this work at a local scale, Mjøsbåtarkivet has published records of 102 small open clinker-built boats from Lake Mjøsa (Mjøsbåtarkivet, n.d.). Boats selected for digitization are chosen based on age, builder, hull characteristics, oral histories, and practical accessibility. The resulting 3D models will be registered in the DigitaltMuseum catalogue and, together with archival and owner information, will provide a representative overview of this regional boat type (DigitaltMuseum, n.d.). These clinker boats present an ideal testbed for developing advanced 3D documentation workflows. Their construction involves overlapping planks, subtle variations in joinery, and distinctive features such as outriggers—all of which challenge conventional surface-scanning techniques. Moreover, the boats embody both tangible craftsmanship and intangible cultural heritage, now formally recognized through UNESCO’s inscription of Nordic clinker boat traditions (UNESCO, 2021). By experimenting with parametric reconstructions of these vessels, the project not only refines digital documentation methods but also contributes to safeguarding boatbuilding knowledge, ensuring that structural principles as well as external forms are preserved for future research, education, and community engagement. Materials and Methods Photogrammetry Because documentation had to take place under varying conditions outside the museum, photogrammetry was chosen as the primary technique. The boats feature diverse surfaces—from white paint to untreated wood—and fluctuating daylight or artificial light placed high demands on the photographer’s skill. To build competence in photogrammetry and test the applicability of 3D data for replica construction, a 1946 boat by renowned builder Martin Grøndal (1891–1973) was selected. The vessel was preserved by the Østby family until 2022 and then given to Hamar Roklub 1881. The vessel was stabilized 50 cm above the ground to allow access to interior and exterior surfaces as much as possible. Loose parts like oars, oar locks, seats and floor was removed in favour of documenting the geometry of the hull. Using a handheld camera, 1,500–2,000 images were captured. Images were first adjusted in Adobe Photoshop for lighting consistency, then processed in RealityCapture. The resulting models displayed imperfections due to rain and other disturbances. Three photo sessions and three 3D reconstructions were attempted, but none produced a fully satisfactory result. Various correction techniques were tested, though with limited success. Aligning 3D Models with Conventional Documentation Care was taken not to artificially “correct” the geometry of the original boat—for example, by straightening the keel in Rhino before generating 2D drawings. Instead, natural asymmetries and deformations were preserved, reflecting the lived history of the vessel. Traditional documentation methods, by contrast, often idealize boats: measurements are taken from one side (usually starboard) and mirrored, erasing differences between port and starboard (Christensen, 1992). Such conventions, rooted in 18th-century Royal Navy practices, produce abstracted drawings rather than faithful records of workboats (Leijonhufvud, 2022). This tension between idealized and empirical documentation guided our approach. We compared 3D outputs with archival drawings (e.g., scale 1:10 or 1:20 plans in the Norwegian Maritime Museum), aiming for compatibility across formats while recognizing that new and old documentation styles embody different conventions. Producing Data for Replica Building The RealityCapture model was rendered in Blender and then used to create line drawings in Illustrator, referencing Christensen’s conventional drawings of similar boats (Christensen, 2013). Renders were produced to highlight starboard views, top views, longitudinal cross sections, and frame placements. These provided the boatbuilder with essential reference points for molds, frames, and planking. Cross sections, though challenging due to surface imperfections, were central to replicating geometry. To minimize interpretive bias, raw 3D cross-section renders were printed at 1:1 scale for direct use by the boatbuilder. These prints indicated plank overlaps, keel placement, and key reference lines. By flipping paper templates, the builder could compare starboard and port variations and settle on practical compromises. Figure 4 - Constructing a replica using 1:1 plots. During construction, the boatbuilder used thin mock-up planks to test curvature, particularly for the garboard plank, which required steaming and bending to extreme angles. While 3D-derived points offered crucial guidance, final shaping decisions relied on the builder’s expertise and “eye” for fair curves. The digital data thus provided essential geometric anchors but not a complete blueprint. Additional information—such as dimensions of screws, nails, and rivets, as well as undocumented loose parts like oars and floorboards—had to be obtained directly from the original boat. The boatbuilder also identified the planking as pine by grinding through the interior paint, which contradicted the common view that spruce was the standard material for these boats. Later, the previous owner confirmed the vessel had undergone repairs, raising the possibility that pine had replaced the original spruce in some sections. Reverse Engineering and Procedural Modeling In parallel with the physical reconstruction, a digital tool was developed to formalize clinker boatbuilding rules. Reverse engineering—defined as reasoning backward from an artifact to its underlying design principles—was applied to translate construction knowledge into an HDA. Figure 5 - Screenshot of the HDA: left, scan overlaid with parametric model; right, parameter interface and node network. This approach aligns with what Leijonhufvud (2022) describes as a forensic method of craft documentation, which combines sensory experience, material study, and digital recording in reconstruction processes. While his work emphasizes reconstruction through physical craftsmanship, our reverse-engineering workflow translates key construction principles into procedural logic. In this way, the Houdini Digital Asset emulates the reasoning of a boatbuilder engaged in manual reconstruction, enabling iterative interpretation within a computational framework. The core functionality of the tool is to slice the 3D scan into sections, visualize these cross-sections, and use them as anchors for defining plank geometry. Using these anchor points and adjustable parameters such as plank thickness, width, and overlap, the tool generates clean geometry that follows predefined construction rules. The modeled planks can then be overlaid on the scan with variable transparency for visual comparison and refinement. Olstad F (2022). Da sporten erobret Norge. Kjønn, klasse, nasjon, modernitet. Pritchard D, Rigauts T, Ripanti F, Ioannides M, Brumana R, Davies R, Avouri E, Cliffen H, Joncic N, Osti G, Toumpouri M (2021) Study On Quality In 3D Digitisation of Tangible Cultural Heritage. https://doi.org/10.4995/arqueologica9.2021.12113. Skjærbakken, E. (2005). Indre østlandets trebåter: langs Glommavassdraget og østlige deler av Drammensvassdraget. Norsk Skogmuseum. Suarez M, Parke F, Castro F (2019) A Procedural Approach to Computer-Aided Modelling in Nautical Archaeology. 3D Recording and Interpretation for Maritime Archaeology, Coastal Research Library, 31. https://doi.org/10.1007/978-3-030-03635-5_8. Tanner P (2013) 3D Laser Scanning for the Digital Reconstruction and Analysis of a 16th century Clinker Built Sailing Vessel. ACUA Underwater Archaeology Proceedings 2013, 137-149. Tanner P (2017) The testing and analysis of Hypothetical Ship Reconstructions. Baltic and Beyond, Change and continuity in Shipbuilding, Proceedings of the Fourteenth International Symposium on Boat and Ship Archaeology, 171-178. UNESCO (2021) Nordic Clinker Boat Traditions. https://ich.unesco.org/en/RL/nordic-clinker-boat-traditions-01686. Várady T, Martin R R, Cox J (1997) Reverse engineering of geometric models—an introduction. Computer-Aided Design, 29, 4, 255-268. https://doi.org/10.1016/S0010-4485(96)00054-1. Wilkinson M, Dumontier M, Aalbersberg I et al (2016) The FAIR Guiding Principles for scientific data management and stewardship. Sci Data 3, 160018. https://doi.org/10.1038/sdata.2016.18. Yamafune K, Torres R, Castro F (2017) Multi-Image Photogrammetry to Record and Reconstruct Underwater Shipwreck Sites. J Archaeol Method Theory 24, 703–725. https://doi.org/10.1007/s10816-016-9283-1.