Using Satellite Imagery to Combat Illegal Fishing

Researchers estimate that the global economic impact of illegal, unreported and unregulated (IUU) fishing exceeds $20 billion annually. This amount represents lost tax revenues, lost economic opportunities, lost jobs and – above all – an expanding, dangerous black market exploiting human and natural resources.

To fight this form of organized crime, ImageSat International (ISI) developed Kingfisher – a multi-sensor, multi-layer maritime intelligence system combining a variety of information sources to expose the covert movements of IUU fishing fleets. Utilizing Satellite Automatic Identification system (S-AIS) data, Synthetic Aperture Radar satellite imagery, electro-optical satellite imagery, Vessel Monitoring Systems (VMS), coastal radar, open source intelligence (OSINT) and even weather patterns, the system goes further than others in detecting the tell-tale behavior patterns of IUU fishing.

One of the system’s major advantages is its ability to capture and analyze satellite images of suspect activities. Satellite detection of moving objects in marine environments is often considered equivalent to finding a needle in a haystack. Kingfisher changes all of this: the system's deep learning algorithms can identify vessels in existing satellite imagery, and by correlating this with other sensor data, they can automatically detect non-cooperative vessels whose AIS is turned off.

Kingfisher also adds an artificial intelligence (AI) twist: a behavior pattern recognition and prediction, multi-agent based simulation integrated with deep learning techniques. This AI enables users to improve satellite usage for rapid tasking, leading to lower costs and greater effectiveness.

Case Study

One project between ISI and the coast guard of a South American country proved the law enforcement potential of the Kingfisher System. In one particular example, the joint team monitored a fishing vessel that approached the country’s Exclusive Economic Zones (EEZ) and turned off its AIS system. The vessel sailed on the outside edge of the EEZ for a time, turned into the EEZ for a few hours, then sailed out of the country’s national waters.

Under ordinary circumstances the ship might have slipped away. But Kingfisher's predictive behavior algorithm forecast the vessel’s course, and it came up with a result that differed from the predictions of the coast guard team’s existing system. Based on Kingfisher’s recommendation, the coast guard dispatched units to intercept, and within a short time, they reported engagement with the vessel at precisely the point calculated by the AI algorithm. The ship and its crew were brought to shore to face charges and the fish haul was confiscated.

This example shows the system's value for counter-IUU fishing, but it is just as useful for identifying vessel movements related to piracy and terrorism, drug smuggling, human trafficking and illegal immigration. Its multi-sensor, multi-layered system delivers analysis from the initial stages of strategic surveillance up to the profiling of individual vessels.