An Artificial Intelligence-Based Indoor 3D Modeling System for Intelligent Spatial Reconstruction and Visualization

Abstract

Indoor 3D modeling has become an important technology for smart building management, digital twin construction, indoor navigation, facility maintenance, and virtual reality applications. However, traditional indoor 3D model construction usually requires professional equipment, manual measurement, and time-consuming modeling procedures. These limitations increase labor costs and reduce the efficiency of indoor spatial digitization. To address these challenges, this study proposes an artificial intelligence-based indoor 3D modeling system that integrates image acquisition, computer vision, depth estimation, semantic segmentation, spatial reconstruction, and interactive visualization. The proposed system uses indoor images or video sequences as input and automatically extracts spatial features, recognizes indoor objects, estimates scene depth, and generates a structured 3D indoor model. By combining artificial intelligence and 3D reconstruction techniques, the system can reduce manual modeling workload and improve the efficiency of indoor environment digitization. The proposed framework can be applied to smart campuses, factories, offices, museums, hospitals, exhibition spaces, and indoor facility management. The results are expected to demonstrate that AI-based indoor 3D modeling offers a practical, scalable solution for intelligent spatial reconstruction and digital twin applications.