Patch-Wise Analysis and Reconstruction of Large-Scale Floorplan for Digital Twin Modeling

Abstract

Transforming 2D floorplans into 3D models is a crucial step toward building digital twin environments. However, existing methods designed for residential floorplans often suffer from errors caused by varying wall thickness when applied to large layouts and rely on strict orthogonality assumptions, limiting their applicability to complex designs. To overcome these limitations, we propose a patch-wise analysis framework that handles both residential and commercial layouts. Our approach leverages a modular segmentation-and-vectorization pipeline, enabling robust and accurate reconstruction of detailed 3D geometry. Experimental results demonstrate that the proposed method achieves superior performance compared to existing techniques, effectively generalizing from residential to complex large-scale floorplans. Our framework thus provides a scalable solution for creating digital twins of diverse real-world environments.