Making Fisheries Monitoring Accessible and Scalable With Alexander Dungate

“One fish, two fish, red fish, blue fish.” You may know this children’s book, but did you know that video footage from fishing vessels is manually reviewed in a similar way? That’s right, typically, a human is required to review hours of fishing trip footage (from vessels that have on-deck cameras installed). But OnDeck Fisheries AI is tackling this challenge with its automated catch monitoring platform that aims to make review faster and more affordable.

The Ocean Impact Podcast · Making Fisheries Monitoring Accessible and Scalable With Alexander Dungate

OnDeck won the Monitoring Spotlight Award presented by Auxillium Foundation at the Ocean Impact Pitchfest last year, and is a two-time Pitchfest finalist for both 2022 and 2023. In this episode of the podcast, we chat to CEO and co-founder Alexander Dungate about OnDeck being awarded funding through Canada’s Ocean Supercluster for the AI for Scalable Fisheries Monitoring Project. And how their dynamic team is working on scalable monitoring solutions that can be deployed worldwide, as well as some of its real-world benefits. Including data sovereignty for Indigenous Nations, enhancing seafood traceability, enabling sustainable fisheries management, and much more.

See below for show notes & relevant links!

Show Notes

share your spiel about the ocean challenge that you're addressing and the wonderful solution you've developed with OnDeck Fisheries AI.

In a nutshell, we're here to make fisheries monitoring accessible and scalable around the world. Fisheries monitoring is the process of seeing what and where fish are being caught out at sea. So as fishermen are out fishing, pulling stuff onto deck, there's security cameras installed on the boat that film what's happening. They come back to shore with thousands of hours of video footage – fishing trips are quite long often – and then back on shore, some poor soul sits down at a computer to watch those thousands of hours of video to manually count all the fish, going one fish, two fish, red fish, blue fish, which is crazy! It's super slow, expensive, and it can't be scaled up around the world. And so at OnDeck, we use AI to automate that video review, making it faster and more affordable to monitor fisheries around the world.

How does the solution come into existing settings? give us a real step-by-step process to fully understand how you plug in and make this positive impact to the ocean.

So I think a great case study for this will be one of our first customers on the west coast of Canada here, Ha'oom. It's a group of five Indigenous nations that are responsible for managing their own fisheries. But they also have oversight from the federal Department of Fisheries and Oceans, the DFO. And so right now, they have these cameras on their boats. Legally, they have to have monitoring in order to go out fishing.

But…the boats go out and they're being filmed as they're fishing. And when the boats come back, the monitoring check that happens is that people watch clips of the video – or all of it, depending on the fishery – just to make sure people were following the rules. And all that information gets sent off to a black box in the federal government. And the Indigenous nations don't have any access to their own data that was collected about their fishing. About is there any bycatch? Are they catching endangered species that they don't want to? Where are their target species? How much are they fishing? All of this critical fisheries management information disappears.

And so, in this case, we're coming in and these cameras are already on the boats. When the boat gets back to shore, we’re able to upload that video footage to the cloud and run it through algorithms that automatically reveal every time a fish was caught, every time a fish was thrown overboard, and things like that. And even from that basic “where was a fish caught? where was a fish discarded?” you can build fishing pressure maps automatically and in nearly real-time, which is a capacity that's never been possible around the world before. And so with that, we're able to help them generate these kinds of compliance checks much more quickly, but also – and most importantly – give them access to and sovereignty over their own data.

This idea of data sovereignty for Indigenous Nations is really cool. That's a really interesting angle. So is that a fishery THAT’s proactive in wanting to adopt a solution like yours? would there have been a requirement placed upon them or did they choose to work with you?

So it is required that they have cameras on board. In almost all Canadian fisheries on the west coast, it's mandatory to have this camera on board, and on the east coast that's starting to to pop up now as well. So it's already mandated that there be this this video review process. But they're choosing to work with us because we can do it way more quickly, way more affordably. And on top of just beating the manual methods, we can give them all this information that's never been possible to collect before around where their fish are, where their fishers are fishing, where the fishing pressures are.

And a real-world reason that they need this data is: say the federal government comes in. Usually fisheries closures happen in massive tranches. Like, “hey, no one's allowed to fish in this massive 100-kilometer area because we're worried about catching yellow-eyed rockfish,” for example. But now these communities, like Ha'oom, have the data to say, “actually, we've been fishing these areas every year and we've never been catching yellow-eyed rockfish. We'd like to continue fishing in these areas.”

And that's really important for them, to be able to catch fish for even just food. There's a lot of subsistence fishing communities on the west coast of Canada. There's no grocery stores within any drive, it’s boat access only. So being able to continue fishing is really important. So giving them access to data to be able to speak to sustainable fisheries management is something they've never been able to do before. So that's really cool.