The U.S. Coast Guard’s (USCG) icebreaker Healy has embarked on a pioneering three-month scientific expedition in the Arctic. An onboard infrared imaging system will amass a unique dataset, aimed at honing artificial intelligence (AI) analysis tools.
Collecting Data for AI Research and Maritime Safety
The Healy’s journey across the North Pole will not only amass critical data but will also pave the way for safer maritime navigation and strengthen our national defense by broadening maritime domain awareness. The work is led by the AI Software Architectures and Algorithms Group at Lincoln Laboratory, a researcher Jo Kurucar commented on the potential impact of this mission.
The diminishing Arctic ice due to global warming is making the area more accessible to both military vessels and illegal fishing activities, posing security challenges for the United States. The evolving Arctic climate and its effects on local wildlife and geography also raise various environmental concerns.
Expanding Arctic Imagery Datasets
Currently, there is a dearth of extensive Arctic imagery datasets for research. While satellites and aircraft provide overhead images, they offer limited insights. An outward-facing camera on a ship, like the one aboard the Healy, can yield more comprehensive images, offering different perspectives of objects such as other vessels. The collected data can then be utilized to train AI computer-vision tools, assisting the USCG in planning naval missions and automating analysis.
Leveraging Advanced Tech for Data Collection
The Healy, being USCG’s largest and most technologically advanced icebreaker, is aptly equipped with a state-of-the-art sensor to collect this dataset. The lab research team collaborated with the USCG Research and Development Center to create the Cold Region Imaging and Surveillance Platform (CRISP). This platform includes a long-wave infrared camera from Teledyne FLIR, which is specially designed for harsh maritime conditions. The camera, capable of stabilizing itself during rough seas and imaging in complete darkness, fog, and glare, is paired with a GPS-enabled time-synchronized clock and a network video recorder.
Public Release of the Dataset
The CRISP project team is set to publicly release the 4-terabyte dataset after the conclusion of the USCG’s scientific mission. The objective behind this release is to enable the wider research community to develop better tools for Arctic operations. In addition, the lab team also plans to provide a baseline object-detection model to aid further research.
Future Impact of the Dataset
This project not only assists with USCG missions but can also be instrumental for researchers applying AI to Arctic data for climate change combat. Paul Metzger, who leads the AI Software Architectures and Algorithms Group, is optimistic about the advances from employing AI to address national challenges.
Related Articles
