3DMGV
A Linux-first toolchain to convert videos into textured 3D models using OpenMVG, OpenMVS, and Python automation scripts.
What it is
3DMGV automates a video-to-3D workflow: extract frames, run Structure-from-Motion (OpenMVG) to estimate cameras and a sparse point cloud, then densify and reconstruct textured meshes with OpenMVS.
Features
- Extract frames from videos (frame sampling + optional background removal)
- Structure-from-Motion pipeline (OpenMVG) for camera estimation and sparse reconstruction
- Multi-View Stereo (OpenMVS) for densification, mesh reconstruction, and texturing
- Python runner that orchestrates the end-to-end workflow and utility scripts
Planned
- Make background removal optional in the pipeline
- Add parameter presets for quality vs speed
- Improve robustness for low-quality / variable frame-rate videos
- Better output organization and cleanup
Technologies
- Python 3
- OpenMVG (SfM)
- OpenMVS (MVS / mesh / textures)
- OpenCV
- rembg
Screenshots
