Objective
Identify bird species on embedded hardware for population monitoring. The project pairs computer vision with a Raspberry Pi 5 to support observation outside a desktop computing environment.
AI · Computer Vision
Intelligent Bird Recognition System
Built an OpenCV and YOLOv11s pipeline on Raspberry Pi 5 using 2,000+ annotated images. The system recorded 15 FPS and 87% recognition accuracy across four species.
Development process
Data & model
Collected and annotated more than 2,000 images for a YOLOv11s bird-recognition pipeline on Raspberry Pi 5.
Embedded optimisation
Selected YOLOv11s for deployment on an 8 GB Raspberry Pi 5, using rpicam for capture and CSV files to record recognition data.
Reported performance
The current homepage records 87% species classification accuracy and 15 frames per second on embedded hardware.
Photos & drawings
Build photos and drawings will be added here as the project documentation is assembled.
Video demonstrations
Project footage is not available yet. Demonstrations and process walkthroughs will be added here.