Fauna-Eye is a wildlife observation system for my backyard. It detects movement, captures images, identifies animal species, and sends notifications.
Once, in the mid of the night, I saw a fox staring trough our terrace toor for a brief moment before it rushed off. When I told my wife the next day, we got very curious which animals visit our garden (everyone is welcome, except snails and mosquitoes). I then came up with the plan to build my own wildlife observation system.
Note
This is a hobby project built primarily for learning and experimentation. I'm not optimizing for simplicity, but for experience and my personal satisfaction with the implementation. As a result, some parts are itentionally more complex than necessary. I still try to keep the code maintainable.
- Setup Hardware
- Capture Image with the camera
- Trigger image from motion sensor
- Provide upload service
- Transmit image to server
- Queue uploads when not connected
- Persist images and metadata
- WebUI to show and manage images
- Send notifications
- Identify species using WEB-API
- Identify species using local model
- Use species information to filter notifications
- Show species information in WebUI
- Add a microphone
- Record bird voices
- Recognize bird species (BirdNET)
- Extend notifications
- Build a proper hardware case
develop: Compose files to build and run components locally. docs: Contains documentation. server: Compose file, dotenv and configs of the server. upload-service: A HTTP upload service with REST API. wildcam: The application running on the outdoor camera.
- Checkout this repository
- In develop/ create a .env file and add your PUID and PGID.
docker compose -f develop/compose.yml up
This includes a wildcam service that generates artificial images.
- Enable SPI
- Checkout this repository or copy the wildcam onto the Pi
sudo apt install python3-picamera2 python3-gpiozero python3-luma.oled python3-prompt-toolkitcd wildcam && python -m wildcam