Building blocks for 3D fuelbeds: object-centered scanning and meshing protocol

Abstract

Accurate modeling of wildland fuelbeds requires knowledge of not only where fuels are in three-dimensional (3D) space but also what they are. In this study, we introduce an object-based scanning protocol designed to generate detailed three-dimensional mesh models of individual fuel particles (e.g., seedlings, shrubs, litter, and cones) using an industrial-grade laser scanner. While traditional terrestrial laser scanning (TLS) or photogrammetric approaches tend to require objects to be segmented from broader-scope environment-level point clouds, our approach begins with the object itself. By scanning discrete plant parts in controlled conditions and capturing their morphology, surface area, and volume at sub-millimeter precision, we create a methodological foundation for fuel characterization that is structurally explicit and ecologically specific. We also propose a flexible workflow to adapt the scanning process for the extensive natural range of variation in fuel object structures, classifying individual objects based on their structural complexity. Digital twins of wildland fuel plants and particles serve as building blocks for future integration with machine learning techniques to improve wildland fuelbed classification and simulation. Our approach shifts the basis of 3D fuels modeling from environmental scanning toward object-driven understanding with implications for fire behavior, emissions, and ecological modeling.