3D Reconstruction Systems in Practice

Signature Projects Featured

(2024–2025) Mobility One Inc.

Online/Offline Real-Time 3D Reconstruction Using Drone Imagery

Research on real-time and offline 3D reconstruction pipelines utilizing drone-captured imagery.

• Online modeling with SLAM + MVS integration
• Offline high-quality reconstruction with extended processing
• Optimized for aerial surveying and robotics

Real-time SLAM MVS Gaussian Splatting

(2025) Mobility One Inc.

3D Reconstruction with Diffusion-based Inpainting

A reconstruction pipeline that enhances drone-based 3D models by filling unreconstructed regions using diffusion models.

• Sparse structure from motion with COLMAP
• Dense mesh & texture generation with OpenMVS
• Diffusion-based inpainting for missing areas

SfM MVS Diffusion

Project List Industry & Research

• Defect detection analyzing multi-focus (Top/Bottom) imagery of identical samples(PoC), Daeduck Electronics, Korea (Mar. 2023 – Apr. 2023)
• Development of a defect object detection model and equipment integration via C++ SDK, AI Voucher Project, Korea (May 2023 – Dec. 2023)
• Defect detection across multi-angle imagery of identical samples(PoC), Daeduck Electronics, Korea (Apr. 2023 – Jun. 2023)
• High-resolution cosmetic container defect inspection, Seobaujeok, Korea (Mar. 2023 – Dec. 2023)
• Online/Offline Real-Time 3D Reconstruction Using Drone Imagery, Mobility One Inc., Korea (Oct. 2024 – Apr. 2025)
• 3D Reconstruction from Drone Imagery with Diffusion-based Inpainting of Unreconstructed Regions, Mobility One Inc., Korea (Jun. 2025 – Dec. 2025)