Ribbit Radar
Bioacoustic Machine-Learning model fine tuned for the detection of ecologically significant frog species
Motivation
Built for biologists, Ribbit Radar significantly reduces the time spent analyzing audio data from the field. Initially developed to aid in monitoring the endangered California Red-Legged Frog as it is reintroduced to ponds in Southern California, Ribbit Radar was later expanded to identify the invasive American Bullfrog, a major threat to native biodiversity.
Function
Given a folder of audio data, Ribbit Radar autonomously processes all files, generating a spreadsheet with timestamps of identified species of interest. It includes metadata such as recording device ID, date/time, and air temperature. The tool supports user inputs for inference settings and output formatting.
My GitHub repository contains the open source code, as well as more technical details.
Continued Development
With robust monitoring in place for the California Red-Legged Frog, our focus has shifted to preventing American Bullfrog invasions. Rapid intervention is crucial, and this requires real-time monitoring solutions.
To meet this need, we are developing an autonomous real-time monitoring system capable of remotely alerting biologists to bullfrog vocalizations detected in the field.
Technical Documentation
For comprehensive technical details on acoustic frog monitoring methodology, deployment guides, and best practices, explore our detailed documentation on GitBook.
Made in Collaboration With
