BirdCam - A real-time garden bird classification with Jetson Nano

Hello everyone,

I would like to share a fun project that I did. I have a bird feeder in my garden and become interested in learning more about the birds that come to this feeder, so I developed a system to keep track of the animals that visit my feeder.

BirdCam is a framework on Jetson Nano for real-time classification of urban faunas. The framework utilizes both traditional image processing and a machine learning approach to perform real-time classification. It has a relatively simple hardware setup with only a Jetson Nano and a compatible camera (with an optional LED illuminator).

Here is the GitHub repository for the project: plertvilai/birdCam_jetson: BirdCam with Jetson Nano (github.com)
There are more information on the GitHub page.

In addition to collecting data of birds, the framework can be used to react to unwanted visitors in real time. I programmed the Jetson Nano to control a simple water pump that shot a stream of water when rats visited the feeder. This temporarily helped deter unwanted visitors to the bird feeder although the rats were quick to realize that food was worth getting wet.

Hope you enjoy the project!

3 Likes

Clever project. Thanks for sharing with us! What’s the most unusual bird you’ve seen at the feeder?

1 Like

Thanks a lot!
I wouldn’t say rare, but I can notice migrating birds with the feeder. California towhees are in my area mainly in spring, and that is when they are showing up on the camera!

1 Like