- Google developed SurfPerch AI tool to assess coral reef health.
- It was trained on audio recordings of coral reefs.
- Incorporating bird sounds boosted SurfPerch’s performance in monitoring reefs.
SurfPerch tunes in to underwater symphony
Google has developed a cutting-edge AI tool, SurfPerch, to aid marine biologists in assessing coral reef ecosystems’ health.
Jointly created by Google Research and DeepMind, SurfPerch was trained on thousands of hours of audio recordings from coral reefs, enabling scientists to “hear reef health from the inside.”
This innovative tool can track reef activity at night and in deep or murky waters, areas previously challenging to monitor.
To fine-tune SurfPerch, Google crowdsourced data through the “Calling in Our Corals” website, where visitors listened to reef sounds and identified fish vocalizations.
This approach yielded a comprehensive “bioacoustic” dataset focused on reef health.
Making waves: AI learns from feathered friends
Remarkably, researchers discovered that incorporating bird recordings boosted SurfPerch’s performance, as common patterns exist between bird songs and fish sounds.
Initial trials utilizing SurfPerch and the crowdsourced data revealed differences between protected and unprotected reefs in the Philippines, tracked restoration outcomes in Indonesia, and provided insights into fish communities on the Great Barrier Reef.
