A lightweight application uses artificial intelligence and a Raspberry Pi to identify bird species from their calls, offering a low-cost and privacy-friendly solution for bird monitoring.
BirdNET-Pi is an open-source application that transforms a Raspberry Pi into an automated bird monitoring system capable of recognising bird species by analysing their songs and calls in real time. Designed to run locally on affordable hardware, it continuously records audio from a connected microphone and identifies nearby birds without requiring cloud-based processing.
The software uses the BirdNET deep learning framework to classify the sounds made by birds, save audio clips, and display them using the internal web interface. The software allows browsing through bird detections, analysing audio recordings, studying bird activity patterns, and visualising the bird observations gathered over time.
It is also possible to use BirdNET-Pi in conjunction with community websites such as BirdWeather, where one can share the bird observations with other users without losing control over the monitoring process. The software also supports outdoor deployment.
The system is compatible with different models of Raspberry Pi and requires only a USB microphone or a sound card to operate. Since all processes are handled by the Raspberry Pi, it presents an energy-saving and inexpensive solution compared to other wildlife monitoring systems.
By combining affordable hardware, AI and open-source software, BirdNET-Pi allows amateurs, teachers, and scientists to monitor birds in their area and study ecology.
