“Even if all the nations of the world agree to create large marine protected areas in the high seas, how are you going to enforce this?” says Liu. “The policy can’t just stay on paper, you need to have practical ways to make sure it’s going to work.”
In lieu of expanding expensive Coast Guard units, most countries will need to rely on satellite data and monitoring tools from organizations like Global Fishing Watch. But while these programs provide a comprehensive look at maritime traffic, they rely heavily on GPS data from these vessels—those ship identification (AIS) numbers mentioned above. Because of this, not many programs are able to track illegal ships that have “gone dark” by turning off their GPS transponders.
Hope to improve these systems, however, may lie in automation.
At WHOI, Liu and a team of engineers and researchers passionate about mitigating illegal fishing are looking into testing the use of artificial intelligence to fill in the criminal activity gaps that existing vessel monitoring systems can’t. By feeding a computer algorithm with databases of illegal fishing vessel images and descriptions, they hope to teach it to cite illegal vessels through high-resolution satellite feeds in real time.
If it works, she says, they will be able to create a program that can codify the characteristics of illegal vessels, possibly alerting local coast guard patrols with SMS alerts on their phones in the not-too-distant future. To find hope that such a project could work, she adds, you need look any further than the progress of marine mammal tracking tools.
“There are examples of good machine learning everywhere that use satellite images to detect whales in the ocean, which are much smaller than the vessels we’re trying to detect,” says Liu. “I think that’s promising for the work we’re doing.
“I think if we can make this happen, it’s going to change the world.”
An illegal driftnet nearly 8-miles-long seized by the Sicilian Coast Guard. (Photo courtesy of Alessandro Bocconcelli)
The AI-based project referenced in this piece, “Seeing the Unseen: Tools for tracking illegal vessels at sea,” is currently led by a team of private investigators at the Woods Hole Oceanographic Institution, including Yaqin Liu (Assistant Scientist, Marine Policy Center), Tom W. Bell III (Assistant Scientist, Applied Ocean Physics & Engineering), Daniel Zitterbart (Associate Scientist, Applied Ocean Physics & Engineering), Lee Freitag (Principal Engineer, Applied Ocean Physics & Engineering), Austin Greene (Postdoctoral Investigator, Marine Chemistry & Geochemistry), and Alessandro Bocconcelli (Emeritus Research Scholar, Applied Ocean Physics & Engineering).
