Low-cost multi-sensor scanning for as-built digital modelling of irregular bamboo structures in field conditions

This study develops and field-tests an accessible digital modelling workflow for capturing the as-built geometry of irregular bamboo structures using low- and mid-cost scanning tools. Bamboo offers a sustainable construction material, but its geometric variability and craft-based assembly pose challenges for reliable digital documentation and integration into computational design workflows. High-precision scanning systems, while effective, are often impractical in field contexts due to cost, complexity, and deployment constraints. To address this gap, the study evaluates a hybrid multi-device scanning workflow combining a professional mobile LiDAR system (Kaarta Contour), a mid-range RGB-D sensor (Azure Kinect), and a consumer-grade tablet with integrated LiDAR (iPad Pro). These devices were deployed on a full-scale experimental bamboo pavilion under outdoor field conditions. Point cloud data were processed through iterative cleaning, alignment, and selective integration, with the professional LiDAR dataset providing spatial coherence and colour-rich datasets supporting segmentation. The results identify key field-specific challenges, including occlusion, registration drift, segmentation instability, and lighting variability, and demonstrate that a layered, multi-device approach can produce spatially coherent and analytically useful digital models. The study defines realistic performance bounds for low-cost as-built modelling workflows and contributes a transferable method for digital documentation of geometrically irregular, craft-based architectural systems.