← All work
Case study · 02/08
2D → 3D Floor Plans
Automated conversion of 2D floor-plan images into 3D Blender models using YOLO detection.
Scroll to build it
The 2D plan, as the converter reads it.
(01) Problem
Manual creation of 3D models from 2D floor plans was time-consuming. Engineers needed an automated way to extract structural elements and generate geometry.
(02) Approach
- 01Implemented YOLO object detection to identify walls, doors, and windows in input images.
- 02Developed Python scripts to interpret detection outputs and drive Blender API commands.
- 03Automated the extrusion of walls and placement of openings based on detected coordinates.
- 04Integrated a React front end for user interaction and model review.
(03) Architecture
Input ImageYOLO DetectionPython ScriptBlender API3D Model
(04) The hard part
The hardest part was turning noisy detections into clean, connected geometry that Blender could build. This required processing detection outputs to ensure structural continuity before generation.
(05) Result
The system reduced manual modelling time by about 90%.