Orca whales in the path of ships could be injured or killed in an accident.

Southern resident Orcas, also known as killer whales, are an endangered species. Researchers at Simon Fraser University (SFU), Canada, propose to use AI to identify them by their sounds, and then warn approaching ships of their presence. (SFU News)

Hydrophones and machine learning for Orca detection

Ruth Joy, a statistical ecologist and lecturer in SFU’s School of Environmental Science, is leading a project that picks up sounds through a network of hydrophones in the sea, filters whale calls, and immediately use alerts to nearby shipping.

Vessels could slow down or change course and thereby avoid ship strikes that could harm the animals.

The project will use artificial intelligence and machine learning to identify and classify the whale calls. Experts from fields such as biology, statistics, and machine learning will collaborate to develop the system.

Fisheries and Oceans Canada has granted $568,179 in funding. The grant is made under the Oceans Protection Plan–Whale Detection and Collision Avoidance Initiative.

Steven Bergner, a computing science research associate at SFU’s Big Data Hub, is collecting whale call datasets that could run into several terabytes in volume.

The sounds will ‘teach’ a computer to recognize the various species of whale sounds that could be encountered during a day. The system will also learn to distinguish between whale communications and the noise from waves and ships.

Salish Sea Orcas

Joy is already responsible for a project that deployed hydrophone nodes in the Salish Sea to detect Orca whales. Another project uses movable acoustic buoys.

“Southern resident killer whales are an endangered species and people are very fond of these animals,” says Joy. “They want to see that these marine mammals are protected and that we are doing everything that we can to make sure that the Salish Sea is a good home for them.”

Related Story:  The Norwegian AI Lab Helps Salmon Aquaculture

Image credit: Wikimedia Commons