Localizing fish sounds in 3D

​To catalog fish sounds in their environment, we need hydrophones to record sounds and cameras to visually identify the species making them. But what if a fish is making sounds outside the field of view of the camera ? Well, this could be a problem... When we see a fish on the camera, we need to make sure that it is the one making the sounds. This is where the passive acoustic localization comes in.

Imagine yourself in a busy street of a city. Now close you eyes and listen.... The jack hammer across the street from the construction workers, the siren of the ambulance passing by, the bird chirping away above your head, a baby crying in a stroller behind you. Even though you can't see them, you know precisely where all these sounds come from. And all of this is because your ears are few centimeters apart and don't receive the sounds exactly at the same time. Only few microseconds...but this tiny difference is enough for your brain to find out where the sound comes from. For our fish sounds we can use the same principle, except that we use hydrophones instead of our ears and that we are not limited to 2! The more hydrophones you use, the more accurate the localization.

Earlier this year, we published a proof-of-concept paper just showing that. We used data that Rodney had collected several years ago in Cape Cod with 6 hydrophones and a 2 cameras, and were able to localize sounds in 3D and associate them to a fish called the tautog. For our expedition in British Columbia, we will use the same approach. However we can't use the same instruments. The system used in Cape Cod was cabled to a dock and could only record for 2 or 3 hours. For our expedition we need to record for several days at a time and in places that are too far from shore to use cabled instruments. This looks like a detail, but it is actually a fairly big challenge that has kept me busy for some time. Especially for the video camera. There are no video camera systems on the market that can record autonomously for several days (at least at a reasonable price). So, I have been working away on designing a low-cost camera unit based on raspberry Pi computers. First field test to come soon...