A Machine Learning Approach to Identifying and Preventing Illegal Fishing

Abstract

Commercial fishing is the most common type of maritime human activity. Commercial fishing is the for-profit activity of harvesting fish or other marine species using boats, and its effects extend far beyond maintaining ocean health and managing fish populations. The monitoring of marine resources is a critical concern for marine resource management. Illegal fishing operations are prohibited. Overfishing and illegal, unreported, and unregulated (IUU) fishing pose serious threats to the sustainability of fish populations and other marine resources. Commercial fishing industry management has historically been complicated and costly, but technology and openness are bringing about real transformation. Using a variety of machine learning methods, one or more pursuers are trained to successfully catch an evader. This method is used to replicate actual illegal and unreported fishing chase events on both the pursuer and evader vessel agents. The goal of this effort is to reduce and eliminate cross fishing and illegal fishing. The Automatic Identification System, which is gathered by satellites and terrestrial receivers, may detect specific fishing activity based on the movement of vessels over time. Each vessel sends that information out on a regular basis.