arnegiacomo/fugleramme

What it solves

Fugleramme creates a digital bird-watching frame that identifies birds by their songs up to real-time. It transforms raw audio detection into a visual experience by displaying hand-cut 1800s-style natural history illustrations of the detected birds on an e-ink display or a web kiosk.

How it works

The system uses BirdNET-Go to listen to a microphone and classify bird species. Fugleramme polls the BirdNET-Go API to identify which birds are present. It then matches these species to a curated library of over 800 hand-cut illustrations. These images are background-removed and arranged on a textured paper page, with birds sized by their actual body mass and placed according to a layout algorithm that puts larger birds in the center.

Who it’s for

Bird enthusiasts and Raspberry Pi hobbyists who want a local, AI-powered bird detection system that looks like a traditional natural history plate.

Highlights

• Local AI Detection: Uses BirdNET-Go for fully local audio classification.
• Curated Art Library: Includes over 800 cut-outs covering 400+ species, using real public-domain plates rather than AI-generated art.
• E-ink Integration: Optimized for Inky Impression e-ink panels for a paper-like aesthetic.
• Flexible Deployment: Can be run as a systemd service on a Raspberry Pi, in a Docker container, or as a web-only kiosk.
• Scientific Accuracy: Bird illustrations are scaled based on real body mass data from AVONET.