/ 
Every scientific expedition to Antarctica subjects crews to some of the most demanding navigation conditions in the world. In addition to traditional navigation instruments, the Malizia Explorer sailboat is equipped with a computer vision system developed by SEA.AI. This collaboration goes beyond the use of detection equipment. It also helps improve the algorithms using data collected in one of the most challenging environments for navigation.
Detecting ice that radar doesn't always see
Radar remains the primary tool for detecting large icebergs. However, medium-sized ice blocks?sometimes called ?growlers??pose a particular threat. Their low height above the water?s surface reduces their radar signature, yet they are still massive enough to seriously damage a ship?s hull.
This is where the SEA.AI system comes in. Using a forward-facing camera and computer vision algorithms, it continuously analyzes the boat?s surroundings to identify floating obstacles and alert the crew.
Captain Jonathan Morice summarizes the value of this assistance as follows:
"We have radar, which is reliable enough for large icebergs. But then there are these medium-sized, very heavy icebergs that can cause a lot of damage. The radar doesn?t always detect them because they don?t protrude much above the water. That?s why we have SEA.AI."
A decision-making tool that complements the onboard instruments
On board the Malizia Explorer, the SEA.AI interface remains displayed during navigation. The system operates alongside the radar, AIS, and other navigation information without replacing them.
The camera continuously transmits images analyzed by artificial intelligence. When an object is identified, it appears on the control screen to alert the navigator. The system has also proven useful for observing marine wildlife. This has allowed crews to detect whales and other marine mammals before they are visually observed.
One of the observations made by the crew concerns the onboard thermal camera. Contrary to popular belief, an iceberg can still be identified using thermal imaging due to the temperature contrast with the surrounding water. This temperature difference makes it possible to distinguish the ice even in the dark.
Jonathan emphasizes this point:
"The camera was very useful during our missions in Antarctica. I was also surprised by the thermal version. Even though we think of icebergs as cold, the temperature difference between them and the ocean really helps us spot them, even at night."
Antarctic missions are also used to test algorithms
The partnership between SEA.AI and Team Malizia dates back to the early years of the company?s development, which was founded in 2018 under the name OSCAR. The campaigns conducted on Boris Herrmann?s sailboats?and now on the Malizia Explorer?enable the collection of real-world imagery under conditions that are rarely accessible.
This data includes iceberg sightings, marine mammal observations, and navigation in fog, at night, or in various sea conditions. It is used to enrich the training datasets used to train computer vision models.
For SEA.AI, this data from operational sources remains essential. Synthetically generated images can supplement training datasets, but they do not fully replicate the complexity of conditions encountered at sea.
A Partnership Focused on Safety at Sea
The partnership between Team Malizia and SEA.AI extends beyond Antarctic expeditions. The tests conducted in these waters are helping to advance systems designed for use on various types of vessels, ranging from recreational sailboats to commercial vessels, as well as research vessels and unmanned ships.
The feedback gathered aboard the Malizia Explorer illustrates an approach based on observing real-world sailing conditions. For developers and sailors alike, each voyage helps deepen their understanding of the limitations of current sensors and improve the tools designed to enhance human vigilance against floating hazards.
