Expanded Vessel Detections with Sentinel-2 Optical Imagery

A digital map, created with a python package, displays a section of the ocean with outlined borders and circles. An inset zooms in on a satellite view of a ship, while icons and navigation menus appear on the right and bottom edges.
Global Fishing Watch map showing optical imagery detections (Sentinel-2) around Tristan da Cunha, matched to AIS identity information and an image thumbnail of a vessel from satellite imagery.

We’re expanding the view of human activity at sea with the launch of vessel detections based on Sentinel-2 optical satellite imagery. This new dataset adds detections from medium-resolution, daylight imagery, significantly increasing coverage and revealing smaller vessels, especially in nearshore waters where activity is often underreported.

Sentinel-2 imagery operates much like aerial photography from space, capturing visible and near-infrared light. Our machine learning models scan these images to identify vessels, estimate their size, speed and direction, and help distinguish true vessels from objects like buoys, infrastructure or debris.

With this release, we’re adding more than just detections. Each detection comes with an image thumbnail, allowing users to visually inspect the vessel in the context of its surroundings.

MCS officers, researchers and other stakeholders can now:

Why it matters

This release opens the door to new insights in enforcement, marine conservation and coastal fisheries. It’s already being used in research exploring nearshore fishing and vessel behavior in areas with limited AIS coverage.

By combining Sentinel-2 detections with existing AIS and radar datasets, stakeholders can build a more complete picture of human activity at sea.

Explore more

Related Updates

Scroll to Top