360 Extractor Pro

High-performance desktop application and command-line tool for 360° video preprocessing. This tool generates optimized datasets for Gaussian Splatting and photogrammetry (COLMAP, RealityScan) by converting equirectangular footage into rectilinear pinhole views and removing operators using AI.

v2.4.1 - Dynamic AI targeting directly via CLI and resolution-adaptive mask dilation.

Key Features

• 360° to Rectilinear: Reproject equirectangular video to pinhole views with configurable FOV and overlap.
• Dual Interface: Graphical UI for ease of use and CLI for automation.
• Advanced Control: Multiple layouts (Ring, Cube Map, Fibonacci), inclination settings, and selective camera extraction.
• AI-Powered: Automatic operator/object removal (supports 80 COCO classes like humans, vehicles, plants) with adjustable confidence, mask inversion, and intelligent motion-based keyframing.
• Metadata Integration: Extract GPS/IMU data (GoPro, Insta360, DJI) and embed into EXIF.
• Quality Control: Automatic blur detection and filtering.

Installation

1. Clone the repository
2. Install dependencies:

   pip install -r requirements.txt

3. Verify environment:

   python3 check_env.py

Quick Start

GUI Mode

Launch the graphical interface for interactive processing:

python3 src/main.py

CLI Mode

Process videos via command line for automation:

python3 src/main.py --input <video_path> --output <output_dir> --interval 1.0

Documentation

For detailed information on configuration and usage, please refer to:

• 🖥️ GUI & Settings Guide: Detailed explanation of all processing parameters, layout modes, AI filters, and JSON configuration.
• ⌨️ CLI Reference: Complete list of command-line arguments, flags, and automation examples.

Author

Nicolas Diolez

License

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0). Required by usage of YOLO26 (Ultralytics). See LICENSE for details.

Credits

Special thanks to Ultralytics (YOLO26), The Qt Company (PySide6), and OpenCV.