Utilizing AI to enhance monitoring of fishing activities via cooperative and non-cooperative methods

Abstract

The fishing industry is vital for global food security, providing nutrient-rich food, supporting livelihoods, and contributing to economic growth. As the global population continues to grow, it is essential to approach fishing in a sustainable way and preserve marine ecosystems. The United Nations has recognized overfishing as a significant issue. For 2017, it was estimated that about 34% of fish stock was overfished worldwide. However, overfishing isn’t the only threat to our oceans. Illegal, unreported, and unregulated (IUU) fishing also poses a major global challenge.
One way to help solve these problems is through effective monitoring of fishing efforts. The two main ways to do so are via cooperative methods, where fishing vessels are responsible for broadcasting their own data, and non-cooperative methods, which use remote sensing technology to monitor vessels without any involvement from the vessels themselves. This project investigates whether fishing operations in the southern Baltic Sea can be effectively monitored using publicly available data. More specifically, it utilizes the Automatic Identification System (AIS), a cooperative method where ships are responsible for broadcasting information on their location and movements via radio signals and Synthetic Aperture Radar (SAR) images which is a form of active remote sensing technique used in non-cooperative monitoring to detect ships from satellites.
This project uses historical AIS data for the whole year of 2018 to demonstrate the effectiveness of AIS in monitoring fishing activities. A machine learning model was used to predict fishing events from individual ship paths. On a larger scale, spatiotemporal analyses were performed which provided insight into fishing trends and patterns throughout the year. To address the shortcomings of AIS, a deep learning model was employed to detect ships on SAR images, which in turn was used to assess the level of AIS uptake among the fishing fleet in the Southern Baltic Sea. The results demonstrate that both AIS and SAR data can be effectively used in the context of monitoring fishing activities. The combined use of these methods revealed that the vast majority of boats do in fact transmit their AIS data in compliance with European regulations.

Popular Abstract

The fishing industry is vital for global food security, providing nutrient-rich food, supporting livelihoods, and contributing to economic growth. As the global population continues to grow, it is essential to approach fishing in a sustainable way and preserve marine ecosystems. The United Nations has recognized overfishing as a significant issue. For 2017, it was estimated that about 34% of fish stock was overfished worldwide. However, overfishing isn’t the only threat to our oceans. Illegal, unreported, and unregulated (IUU) fishing also poses a major global challenge.
Monitoring fishing activities is a key strategy to combat these issues. There are two main ways to monitor fishing: cooperative methods, where fishing vessels broadcast their own data, and non-cooperative methods which use remote sensing technology to track vessels without their involvement.
This project examines whether fishing operations in the southern Baltic Sea can be monitored effectively by using publicly available data. Specifically, it explores the use of the Automatic Identification System (AIS) and Synthetic Aperture Radar (SAR) images. AIS is a cooperative method where ships broadcast their location and movements while SAR is a non-cooperative method that uses radar technology on satellites to take images of the earth's surface.
The project used historical AIS data for 2018 to monitor fishing activities. An AI model was employed to predict fishing events from the paths of individual ships. Additionally, patterns in fishing activities were explored throughout the study area over 2018 using AIS data. Although AIS is shown to be a great tool for monitoring fishing activities, not all ships use the system. To address this limitation, an AI model was fine-tuned to detect ships in satellite images. This project matched boats found on satellite images with AIS signals to estimate how many ships actually use AIS in the study area and therefore assess the effectiveness of AIS in monitoring fishing activities.
The results showed that the vast majority of boats do lawfully broadcast their AIS data, demonstrating that in this area of the world, AIS can be effectively used to monitor fishing activities. This project indicates that SAR data can be used alone to follow fishing efforts, but because SAR images are taken less frequently than AIS signals are updated and AIS messages contain much more information than ship detections on SAR images, AIS is significantly better at monitoring fishing activities. However, AIS data always rely on the cooperation of the vessel while SAR does not.

• author: Ingason, Ingþór
• supervisor: Pengxiang Zhao; Josefine Larsson
• organization: Dept of Physical Geography and Ecosystem Science
• alternative title: AI-Powered Solutions for Better Monitoring of Fishing Activities
• course: NGEM01 20241
• year: 2024
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: Physical Geography and Ecosystem analysis, Remote sensing, Machine Learning, Deep Learning, Synthetic Aperture Radar, Automatic Identification System
• publication/series: Student thesis series INES
• report number: 669
• language: English
• id: 9172899
• date added to LUP: 2024-08-29 10:05:54
• date last changed: 2024-08-29 10:05:54