Methven Track is a Python application that generates 3D camera tracks from video clips. It leverages COLMAP for structure-from-motion reconstruction and exports results in USD (Universal Scene Description) format.

1. Extract frames from your video using FFmpeg.
2. Detect and match features with COLMAP.
3. Reconstruct the camera track in 3D.
4. Export the camera path and point cloud to USD.

1. Create a New Project Choose a source video and a folder to store project files.

2. Adjust Settings Select compute device, camera model, and feature matching type. Default settings usually work great!

3. Run Tracking Click the ▶ Run Tracking button to start reconstruction.

4. Export to USD Once tracking completes, export your 3D reconstruction to a .usd file.

• ⚠️ Orientation and scale may occasionally need manual adjustment.
• 🐛 Logging system doesn’t print all output — run via terminal for full logs.
• 💡 GLOMAP support could be added in future (Python bindings may not exist yet).
• 🙌 Contributions are welcome — whether bug reports, code improvements, or algorithmic suggestions!
• 🎬 Some challenging shots might still require professional camera tracking software.

👉 Get the latest release from the Releases Page.

Created by Christopher Hosken Inspired by Josh Methven

• 📧 [email protected]
• 🔗 LinkedIn: christopher-hosken