
bnww/dart-sense
📦 Open Source Projectbnww
Real-time automatic dart scoring using YOLOv8 computer vision to track dart positions and calculate scores instantly.
Dart Sense represents a practical application of edge-ready computer vision in sports technology. At its core, the project utilizes the YOLOv8n (You Only Look Once) architecture, a high-performance object detection model, to identify dart impacts on a board. The system includes a robust calibration pipeline that maps the physical board coordinates to the digital scoring grid, ensuring accuracy even with varying camera angles.
Key technical features include real-time inference capabilities, allowing the application to process video streams with minimal latency. The Python-based codebase is designed for accessibility, enabling users to set up a scoring station using standard hardware like a smartphone or webcam. By automating the spatial analysis of dart placements, Dart Sense transforms a standard dartboard into a smart, data-driven gaming environment, making it an excellent example of integrating deep learning into consumer-facing hobbyist hardware.
💡Highlights
- ├─YOLOv8n-based dart detection
- ├─Real-time video stream processing
- └─Automated board calibration
🎯For
- ├─Computer Vision Developers
- ├─Sports Tech Enthusiasts
- └─Hobbyist Makers