BirdNET-Pi is an open-source tool that turns a Raspberry Pi into a 24/7 autonomous bird song monitor, analyzing audio, identifying species via machine learning, and storing detections in a local database
. Beginners need a Raspberry Pi (3B+ or 4), a USB microphone, a MicroSD card, and basic knowledge of flashing images using Raspberry Pi Imager. It operates on 64-bit OS and creates a local web interface for viewing data.
This video provides an overview of the BirdNET-Pi software:
58s
Essential Hardware & Setup
Raspberry Pi: A 3B+ or 4B (preferably 4GB+ RAM) is recommended.
Microphone: A USB microphone, such as a directional mic or a simple USB dongle, is necessary for capturing sound.
Storage: A high-quality 16GB+ MicroSD card.
Installation: Use the Raspberry Pi Imager to install the 64-bit Bookworm OS onto the SD card.
Configuration: Click the gear icon in the imager to enable SSH, set a username/password, and configure Wi-Fi.
Installing BirdNET-Pi
Insert the SD card, power up the Raspberry Pi, and connect it to your network.
Open a terminal (or use PuTTY) to SSH into the Pi: ssh username@birdnet.local.
Run the official installation command provided in the GitHub repository.
The installation takes roughly 10–20 minutes and will reboot automatically.
Configuration & Daily Use
Access UI: After reboot, type birdnet.local into a web browser on your computer or phone to access the dashboard.
Location: Navigate to the settings in the web interface to set your exact latitude and longitude for better species identification.
Microphone Setup: Ensure the correct USB audio device is selected in the settings.
Monitoring: The system will start recording and analyzing immediately. You can view spectrograms and species detections in real-time.
Data Sharing: You can optionally connect your findings to BirdWeather to share data with a global community.
Tips for Beginners
Positioning: Place the microphone in a shielded area to reduce wind noise.
Security: Change default passwords to secure your device.
Updates: The system can be updated directly through the web interface tools section.
Troubleshooting: If the web interface is unreachable, check the IP address via your router and ensure the service is running
BirdNET-pi provides actual bird name identifications along with confidence scores, rather than just raw audio traces. It is a real-time, 24/7 automated acoustic identification system that uses machine learning to identify bird species from their songs, chirps, and calls.
Here is a breakdown of what BirdNET-Pi provides:
Species Identification: It analyzes audio data and identifies the bird species present, such as "American Robin" or "Carolina Wren".
Confidence Scores: Every detection comes with a confidence score (percentage) indicating how likely the identification is correct.
Detailed Data: It logs the timestamp, species name, confidence score, and specific audio clips of the bird call.
Visualization: It generates spectrograms (visual representations of sound) for the bird calls, helping you verify the identification.
Dashboard: It provides a local web interface that shows a dashboard of recently detected birds,, complete with photos of the species.
While it does not identify individual birds, it is highly effective at identifying the species, often cataloging thousands of bird songs in a few days.
BirdNET-Pi is actively used in the UK for monitoring and identifying bird species by sound. It is a popular open-source, DIY project run on a Raspberry Pi, with numerous users in England, Scotland, and the UK generally setting them up in gardens and, in some cases, using them for scientific research.
Here are specific details regarding its usage in the UK:
Active Stations: BirdWeather maps show multiple BirdNET-Pi stations operating across the UK, including locations like Blackpool and Leek.
Active Research/Projects: It is being used for serious monitoring, such as in a project for the re-introduction of Turtle Doves in Southeast England.
User Experiences: UK-based users have reported setting up systems, including placing microphones in roofs to monitor gardens, noting it works well for detecting a range of residential, farmland, and migratory species.
Configuration for UK Species: Users in the UK have shared tips for configuring the software, such as removing prefixes like "European" or "Eurasian" from common bird names for better local reporting.
Community & Discussions: UK birdwatchers discuss their results on forums, noting the system's ability to identify species like redwings and other common garden birds.
The software is highly regarded for being able to run offline, making it a popular, low-maintenance option for continuous, 24/7, local monitoring of bird
While there is no single public directory of all BirdNET-Pi users in the Highlands, the region's unique biodiversity makes it a prime location for acoustic monitoring. Projects often range from individual "citizen science" setups to formal academic research.
Academic and Research Projects
Universities in and around the Highlands are increasingly using acoustic AI for wildlife monitoring:
University of the Highlands and Islands (UHI): The Environmental Research Institute (ERI) at UHI North, West, and Hebrides (Thurso) conducts extensive avian research. They currently lead projects involving semi-automated analysis and citizen-science coordination, which often align with BirdNET-Pi's capabilities. You can contact Dr Neil James (neil.james@uhi.ac.uk) for insights into regional bioacoustics projects.
Cairngorms Connect: This massive landscape restoration project in the Highlands uses various monitoring tools to track species recovery, including predators and their prey. Their science and monitoring team often explores efficient field sampling protocols.
UCL (University College London): While not based in the Highlands, their People and Nature Lab has held workshops specifically on building BirdNET-Pi units and may have collaborators working in Scottish remote regions.
Where to Find and Connect with Users
To get advice from active users in Scotland, these platforms are your best bet:
BirdWeather Interactive Map: Many BirdNET-Pi users sync their data to the BirdWeather Map. You can zoom into the Highlands to see active stations and often find the names or handles of the people running them.
BirdNET-Pi GitHub Discussions: This is the primary technical community. Users frequently share their setups and troubleshoot regional-specific issues (like waterproofing for Scottish weather) in the BirdNET-Pi Show and Tell forum.
Highland Environment Forum: This network connects local conservationists and community groups. Their Nature Projects Directory lists various regional monitoring initiatives where you might find technology-focused volunteers.
Common Uses in the Highlands
Users typically deploy BirdNET-Pi for:
Detecting Rare Species: Identifying "elusive" birds like the Scottish Crossbill (the UK's only endemic bird) or nocturnal species like the Eurasian Pygmy Owl that are hard to find via traditional census walks.
Migratory Tracking: Monitoring the arrival and departure of migratory birds across the Highland moorlands and wetlands.
Citizen Science: Contributing local data to global databases to help map bird sensitivities to environmental changes, such as wind farm developments.
Key Considerations for a Permanent Station
Dun Coillich's terrain and climate present specific challenges for a permanent Raspberry Pi-based station:
Power Supply: As a remote "off-grid" site, a permanent BirdNET-Pi unit would likely require a solar panel and battery setup. Research by University of Oxford teams in similar Highland pinewoods has used passive acoustic monitoring with robust field-deployable kits.
Connectivity: BirdNET-Pi usually requires Wi-Fi to process and display real-time data. For a location like Dun Coillich, you may need a 4G/LTE gateway or to configure the unit for "remote/off-grid" mode using an internal Real-Time Clock (RTC) chip to manage data logs without a live connection.
Weatherproofing: Given the Highland weather, housing the Raspberry Pi in an IP-rated enclosure is essential. Look at "outdoor case" designs used for ADS-B or weather stations.