Three individuals and two teams will split $150,000 in prizes following their win in the xView3 competition that the Defense Innovation Unit hosted to seek an open-source algorithm for detecting dark vessels engaged in illegal fishing.
DIU said Monday it partnered with Global Fishing Watch to facilitate the three-month contest, challenging 1,900 participants in 67 countries to develop the sought machine learning models.
Participants made models in an effort to help users discern between fishing vessels, non-fishing vessels and fixed infrastructure in a large, open-source dataset of maritime radar imagery. The dataset consisted of maritime images taken in day and night conditions.
The algorithms were designed to better analyze data gathered via synthetic aperture radar and identify dark vessels based on the input.
Jared Dunnmon, AI/ML portfolio director at DIU, said the ML models can help U.S. and international partners identify illegal fishing activity at speeds not possible for human analysts.
“Instead of asking human beings to look through specific satellite images we think may be important – which may take several hours per image – we can use modern computer vision algorithms to look through every single satellite image we record in a matter of minutes,” added Dunnmon.
Global Fishing Watch will pay a total of $100,000 to the following winners:
- BloodAxe (Ukraine)
- Kohei (Japan)
- Selim_sef (Belarus)
- Tumen (China)
- Skylight + PRIOR at the Allen Institute for AI (U.S.)
The Skylight team, which is the top U.S.-based prize winner, will receive an additional prize of $50,000 from DIU.