Dutkat: A Privacy-Preserving System for Automatic Catch Documentation and Illegal Activity Detection in the Fishing Industry

Abstract

United Nations' Sustainable Development Goal 14 aims to conserve and sustainably use the oceans and their resources for the benefit of people and the planet. This includes protecting marine ecosystems, preventing pollution, and overfishing, and increasing scientific understanding of the oceans. Achieving this goal will help ensure the health and well-being of marine life and the millions of people who rely on the oceans for their livelihoods. In order to ensure sustainable fishing practices, it is important to have a system in place for automatic catch documentation.

This thesis presents our research on the design and development of Dutkat, a privacy-preserving, edge-based system for catch documentation and detection of illegal activities in the fishing industry. Utilising machine learning techniques, Dutkat can analyse large amounts of data and identify patterns that may indicate illegal activities such as overfishing or illegal discard of catch. Additionally, the system can assist in catch documentation by automating the process of identifying and counting fish species, thus reducing potential human error and increasing efficiency. Specifically, our research has consisted of the development of various components of the Dutkat system, evaluation through experimentation, exploration of existing data, and organization of machine learning competitions. We have also implemented it from a compliance-by-design perspective to ensure that the system is in compliance with data protection laws and regulations such as GDPR. Our goal with Dutkat is to promote sustainable fishing practices, which aligns with the Sustainable Development Goal 14, while simultaneously protecting the privacy and rights of fishing crews.