Paper Contents
Abstract
This paper presents Blueprint3D, an automated system for converting 2D architectural floorplans into structured 3D building models using a hybrid pipeline of image processing, geometric extraction, semantic analysis, and procedural reconstruction. The system addresses the limitations of manual 3D modelingwhich is time-consuming, error-prone, and requires specialized expertiseby detecting walls, rooms, doors, and windows directly from scanned or digital layout images. Techniques such as thresholding, morphological filtering, contour analysis, template matching, and connected-component segmentation are used to extract structural elements, which are then transformed into a semantic representation suitable for automated 3D generation. Using Blenders scripting interface, the system extrudes walls, inserts openings, and assembles complete room meshes into an accurate 3D scene. Evaluation on a diverse dataset of floorplans demonstrates high precision in wall and room detection, reliable geometry reconstruction, and robust export to common 3D formats. Despite challenges with symbol variability and curved structures, the system significantly accelerates architectural visualization workflows. The results show that Blueprint3D provides an efficient, scalable, and accessible foundation for automatic 2D-to-3D architectural conversion.
Copyright
Copyright © 2025 Amol Bhilare, Pranav Patil, Ayush Patle, Varad Patkar, Gaurang Patil. This is an open access article distributed under the Creative Commons Attribution License.
