Algorithms in the Arctic – removing bad weather from images to make Arctic shipping safer

A PhD candidate at the Norwegian University of Science and Technology has developed an algorithm to enhance Arctic shipping navigation by removing weather-induced visual impediments from images, Norwegian High-Tech News reports.

• This innovation by Nabil Panchi aims to make autonomous navigation algorithms more effective in poor weather conditions, similar to how ship captains struggle with limited visibility due to fog, snow, or rain.
• Panchi’s algorithm was trained by using thousands of images from the Arctic, collected during a 2023 voyage of the research ship Kronprins Haakon. The new technology — part of the DigitalSeaIce project — integrates local and regional sea ice forecasting models with shipboard and satellite observations, aiming to build a comprehensive digital infrastructure for Arctic sea ice monitoring.
• The algorithm uses AI to clean up weather-affected images, allowing existing navigation systems to function accurately. Additionally, Panchi and his supervisor, Associate Professor Ekaterina Kim, have published their research and made the SeaIceWeather dataset publicly available. That will provide thousands of clear and weather-degraded images for further study and development.
• The research highlights the increasing importance of safe and efficient navigation in Arctic waters, where shipping activity increased by 25% between 2013 and 2019 due to melting sea ice. Panchi’s work not only aims to improve autonomous shipping technologies, but also to ensure safer navigation through better understanding of ice conditions, ultimately reducing emissions and the risk of environmental disasters.