A groundbreaking new approach from Lancaster University harnesses artificial intelligence (AI) to detect icebergs and floating ice, boasting an impressive 94% success rate. This new AI-powered system is designed to tackle a significant safety concern for the cruise industry and other maritime operations, which navigate through ice-prone regions, offering a much-needed solution to reduce the risks of dangerous iceberg collisions, particularly in polar regions.
As the cruise industry continues to experience growth, surpassing pre-pandemic passenger levels with 31.7 million in 2023, the expedition cruise sector is particularly expanding. With an increasing number of travellers seeking adventurous experiences in remote and icy regions, it has become vital to ensure that safety measures are in place to prevent accidents. Notable incidents, such as the collision between the Carnival Spirit cruise ship and floating ice in Alaska’s Tracy Arm Fjord in September 2024, have highlighted the vulnerabilities in current detection systems. The risk of iceberg and sea ice encounters is ever-present, and while such incidents are rare, those involving smaller floating ice masses, known as growlers, are more frequent and can be more difficult to detect.
The growing demand for expedition cruising has also raised concerns about safety standards, especially as climate change and dynamic itineraries further complicate navigation systems. To address these challenges, Lancaster University’s innovative AI-driven system uses a combination of synthetic aperture radar (SAR) data and deep learning techniques. The system was initially developed for use in astrophysics by Dr. John Stott, a senior lecturer at the university. His research, originally aimed at detecting galaxy clusters in large images of the sky, is now being adapted to identify icebergs and sea ice, offering the potential to transform maritime safety across a variety of sectors.
The AI system operates through region-based convolutional neural networks (CNNs), a form of deep learning that excels at pattern detection and image recognition. This allows the system to identify icebergs and sea ice in satellite radar images, even under challenging conditions like cloud cover or poor lighting, where traditional radar systems often fail. In its trials, the system has achieved a 94% success rate in detecting icebergs in smaller areas of ocean images. For larger regions, such as the Arctic, the accuracy drops slightly to around 80%, but this remains a highly competitive result when compared to current radar technologies.
This new technology is poised to have a significant impact on the maritime industry. Each year, there are typically two or three shipping incidents involving icebergs in the Northern Hemisphere, leading to costly diversions, repairs, and false alarms triggered by radar systems. The AI system offers the potential to reduce these incidents by improving the accuracy and range of iceberg detection, thus lowering the financial and operational impacts on the industry. The technology is also expected to be useful across other maritime sectors, such as fishing and oil and gas, where iceberg detection plays a critical role in reducing ship damage and ensuring safe transport routes.
The project’s development, which has been supported by a £300,000 grant from the UK’s Science and Technology Facilities Council (STFC), is now advancing toward commercialisation. Dr. Stott and his colleague Dr. Sonny Bailey are working to make the AI-powered iceberg detection system a viable product for maritime operators globally. In the next phase of the project, live testing will be carried out to compare the AI’s predictions with real-time conditions experienced by vessels in icy regions. This will allow for further refinement of the system, with the ultimate goal of providing real-time iceberg location data to shipping companies, insurers, fishing fleets, and expedition cruise operators.
This AI-driven solution offers a promising future for the cruise industry and other maritime sectors, helping to mitigate the risks posed by ice hazards and offering a path toward safer, more reliable operations. By reducing the likelihood of collisions and enhancing detection capabilities, the technology could help protect both vessels and passengers while also reducing reputational risks for companies operating in polar regions. With the continued growth of expedition cruising and other maritime industries, this AI-powered iceberg detection system has the potential to play a crucial role in navigating the icy waters of the future.
