Bird activity and environment monitoring system

Inspiration

I was interested in “the types of birds that come around my house” and “the correlation between specific weather conditions and bird activity in specific places (prediction).” This would be very useful for the bird conservation community.

What it does

It detects birds by their songs and automatically collects and compiles both the detected birds and weather information in real-time.

How I built it

Used a Raspberry Pi 4 with BirdNET, an AI-based bird identification system, for bird detection and identification. Used a Raspberry Pi 4 with temperature, humidity, light, pressure sensors, rain gauge, wind vane, and anemometer for weather information (Weather HAT).

Challenges I ran into

Detectability and accuracy for bird identification were the main challenges. I improved these by testing and comparing different microphones and setting a confidence threshold for bird detection.

Accomplishments that we're proud of

I successfully developed a platform to explore the correlation between specific weather conditions and bird activity in specific places.

What I learned

BirdNET is a useful software for detecting and identifying birds from their songs, but there are still challenges in balancing detectability and accuracy for identification.

What's next for my project

• Collect much more data
• Develop dashboards
• Explore potential correlations between bird activity and weather conditions
• Create a prediction model

Built With

• birdnet
• kintone
• python
• raspberry-pi
• sensors