MORRO BAY, Calif. (CN) — The boat gently rocks in the Pacific Ocean on a foggy October morning around 11 miles south of Morro Bay. Then Captain DJ Hardy, standing at the helm with a pair of binoculars, spots significant movement in the water.
On the right side of the ship Freedom, a mother humpback whale and her calf have just appeared above the glassy surface. Like many of his passengers, Hardy grabs a camera — in his case, a Canon R7 with a 400 mm telephoto lens. He zooms in as the mother dives, her dark tail gracefully rising out of the saltwater, and presses the shutter.
The whale’s fluke, or tail, Hardy later explains, is key to identifying these massive creatures of the sea.
“It can be all white, all black, or any combination in between, and they all have a unique shape,” Hardy tells the whale watchers who have joined him on his family-owned Sub Sea Tours and Morro Bay Whale Watching cruise.
Humpback whale populations are making a strong comeback, and scientists are using image recognition technology to identify and track them. That’s created an unlikely synergy between researchers and sightseeing tours like this one.
By submitting their own images of whale tails, non-scientists like Hardy and his passengers have become key contributors to scientific research. “Public participation has been a growing and important part that supplements the work that we do,” said John Calambokidis, a senior research biologist and co-founder of Cascadia Research Collective, a field research nonprofit based in Olympia, WA.
By looking at traits like coloration, markings, scarring and shape, Cascadia was one of the first research organizations to start identifying whales by their tails.
Back then, in the ‘80s, Cascadia took photos of the massive creatures with film cameras. Researchers noted each tail’s unique characteristics, then created what looked like baseball cards for each whale, which were stored in sleeves.
That work “was not too complicated and manageable when we were in the hundreds of whales,” Calambokidis said. Before environmental protections in the early ‘70s, hunting dramatically curbed whale populations and ultimately endangered the humpback, which can grow to 60 feet.
Since Cascadia launched close to 45 years ago, though, humpback whales have prospered. And as populations have rebounded, the work of IDing all of them became more challenging, Calambokidis said.
“Back then, there were about 500 humpback whales using the California coast,” he said. “Now, our estimate is around 5,000.”
Ted Cheeseman, a whale researcher who led wildlife expeditions in Antarctica for 30 years, has also noted the resurgence in whale populations.
Decades of whaling had diminished whale numbers. Even in the 1990s, “there were very, very few whales,” he said.
By 2005, though, Cheeseman began to see whales return to Antarctica. He wanted to share what he was seeing with researchers but wasn’t sure how. “There wasn’t a clear pathway,” he said.
But as technology progressed and whale populations continued to rebound, automated image recognition technology eventually presented a method for identifying large numbers of whales. Machine learning also created opportunities to sort and organize the data.
Image recognition uses artificial intelligence to identify objects, places, people and more.
When someone uploads a photo to Facebook, facial recognition technology suggests the friends it believes are in the photo. The technology has come to serve a range of purposes, including by helping investigators track down people involved in the Jan. 6 riots. And just as facial recognition helps law enforcement find suspects, Cheeseman wondered if it could be used to ID and track whales.
He enlisted the support of Ken Southerland, an aerospace engineer. With input from Silicon Valley experts, they developed the algorithm for whale recognition.
