3D Virtual Modeling Realizations of Building Construction Scenes via Deep Learning Technique

Area of Science:

• Computer Graphics
• Computer Vision
• Artificial Intelligence

Background:

• Traditional architectural drawings require specialized knowledge for interpretation.
• The construction industry increasingly utilizes 3D virtual models for intuitive understanding.
• Manual creation of 3D models from 2D plans is time-consuming.

Purpose of the Study:

• To design and implement an automated system for 3D building reconstruction.
• To convert architectural plans in DXF format into 3D models in 3DS format.
• To improve the efficiency and accessibility of architectural visualization.

Main Methods:

• Analysis of common architectural drawing design rules.
• Development of a system to extract building information from DXF files.
• Application of computer vision, pattern recognition, and AI for 3D model generation.

Main Results:

• Successful automatic generation of 3D models (3DS) from 2D architectural plans (DXF).
• Extraction and analysis of geometric and semantic information from architectural drawings.
• Demonstration of a functional 3D reconstruction system for architectural constructions.

Conclusions:

• Automated 3D reconstruction of architectural drawings is a significant advancement in architectural computer graphics.
• The developed system effectively processes complex architectural data.
• This technology offers a more intuitive and efficient way to represent and understand building designs.