How mwasifanwar/spacesense works
Project type and purpose
• This is an AI-driven backend service named “spacesense” that processes telescope images to detect and track space debris, predicts potential collisions via orbital mechanics, and optimizes satellite trajectories using reinforcement learning. It exposes a server/API entry point (main.py → src/api/server.py) and organizes processing into modular subpackages.
File structure
• Top-level:
• main.py and config.yaml: application entrypoint and global configuration
• requirements.txt: Python dependencies
• src/api: HTTP or RPC server layer
• src/data_processing: ingestion and pre-processing (loading raw imagery, basic transforms)
• src/computer_vision: specialized CV stages (debris_detector.py → tracker.py)
• src/orbital_mechanics: physics-based collision predictor and propagator
• src/reinforcement_learning: agent/environment definitions and training loop
• src/utils/config.py: central config loader
• tests: unit tests for detector and orbital modules
README
• Confirms the multi-stage pipeline: ingest images → detect & track debris → predict collisions → optimize trajectories. No front-end; entirely backend/API with AI/ML & physics components.