“The Sea, once it casts its spell, holds one in its net of wonder forever.”― Jacques Cousteau
This project was originally intended to contribute to OrcaSound, a nonprofit collaborative group in the PNW that I contribute to.
OrcaSound is a collection of audiophones (underwater microphones) coupled to an ML backend that alerts listeners to wale activity. At anytime, users can listen to the current audiophones online. The main focus is tracking the J-Pod, a local group of Orcas as they migrate around the region. The classifier distinguishes non-biologic noise (boats) from other biologic noise (dolphins plus other marine creatures) and focuses on the distinct sounds of the J-Pod, whose clicks and squels are unique, as with other Orca pods.
Current tasks in the project mostly revolve around making the audiophones cheaper and easier for more people to deploy. As an added requirement, these have to have a ground station for power and internet.
The original proposal was more of an autonomous boat, however, that is a significantly more complex task and one that increases the likelihood of losing the hardware altogether.
The Spresense Smart Buoy avoids that all together by being deployable in any location. While it may or may not be able to stream audio at a high quality, it can record on board and send alerts when it notices wale activity with the onboard microphone. This, while not completed yet, can then couple into the Orcasound backend and alert users to listen to the closest audiophone.
However, thanks to the Spresense board and LTE connectivity, the applications for this hardware are just limited by code and what sensors can be added.
Other applications include:
Monitoring shellfish traps
Being a low-wake alert buoy
Issues: still running into problems with microphone and code on device. Battery system and board work great.
In progress.
Comments