Bird Buddy

We’ve discussed many of the other top applications submitted for our Challenge Climate Change contest (a collaboration between us, QuickLogic, and hackster.io).  Next we’d like to discuss one called “BirdBuddy” – a really interesting project and a 2nd place finisher in the contest.

Due to climate change, birds are migrating earlier each year.  Different species in different locations are moving at different times.  This application, designed by Kevin Loeffler, creates a local system that can be trained via AI to identify calls of particular species and so enable amateur “citizen scientist” birders as well as professional ornithologists to track when particular birds arrive and depart the area during their annual spring and fall migrations.  Having a large set of such systems deployed globally could provide a tremendous amount of useful information by tracking and documenting the changing migration patterns of hundreds or even thousands of different species.

The basic building blocks of the system are the SensiML Analytics Toolkit, the QuickLogic QuickFeather Development Kit, an Adafruit solar panel, Li-ion battery and charger, and breadboard and case.

First, Kevin chose two particular species he wanted to track for his area (Southwestern Willow Flycatcher and Whooping Crane), and then found multiple bird call examples for each species (readily available online).  Then he processed the MP3 files (using a software tool called “FFmpeg”) to convert them to 16KHz WAV files ready to be imported into the SensiML Data Capture Lab.  Once those files were imported he was able to use the SensiML tools to quickly and easily create an AI-based model for his system to use to accurately identify when a particular species is calling.

Next, he generated and exported the model file using the QuickFeather board as his target.  With the flash memory on the board loaded with his model, he was ready to test his system.  He connected his solar panel, battery and charging system and then was ready to use the QuickFeather on-board microphone to listen for the birds.  As specific bird calls are identified they are displayed on a host computer connected to the QuickFeather board via a USB port.

In the future, he discussed the potential for transmitting the identified bird data from the BirdBuddy system in someone’s yard via WiFi to a centralized database where the information could be aggregated and analyzed statistically.  Having multiple BirdBuddy systems installed in different locations around the world would make a huge amount of useful (and pre-processed) data available to scientists for further study.  Overall, another great way to collect information to help the world combat the effects of climate change.

A much more detailed description of the project is available on the hackster.io website

See Winning Hackster Line Powered Project Details:

Bird Buddy by Kevin Loeffler