Orca AI: teaching cargo ships to see what radar misses
A Tel Aviv startup mounts thermal and day cameras above the bridge so ships can spot small, unlit vessels that radar and AIS miss, and Lloyd's Register just ran a real Mediterranean trial to check the claim.
I came across Orca AI while reading about a live trial Lloyd's Register ran this spring, and the detail that caught me was how unglamorous the problem is: small fishing boats that radar simply doesn't see.
Founded in 2018 by Yarden Gross and Dor Raviv, a former Israeli navy computer vision specialist, Orca AI builds a bolt-on system for cargo ships called SeaPod. It's a sensor unit mounted above the bridge, combining day and thermal cameras for a full 360 degree view, feeding a computer vision model that detects, classifies, and tracks everything nearby in real time.
The clever part isn't the cameras, it's what they're built to catch. Small, unlit, or low-signature vessels routinely disappear from radar and AIS, the systems ships have relied on for decades. Human watchkeepers get tired during long night crossings through crowded waters. Orca AI's model doesn't replace the crew, it just doesn't blink.
In April, Lloyd's Register put that claim through a real test: five days, 828 nautical miles from Gioia Tauro to Marsaxlokk through some of the Mediterranean's busiest lanes, including the Strait of Messina. Out of 739 relevant targets logged, most of them during congested stretches, SeaPod hit 94% precision and 98.6% recall, and had zero downtime for the whole voyage. More than 1,200 vessels already run the system, with partners like NYK piloting it toward fully autonomous operation.
What stays with me is how mundane the payoff is. Maritime shipping still runs on rules built for eyesight and paper charts, and this small Israeli startup is quietly closing a gap radar left open for decades. No moonshot, just a camera that doesn't get tired at 3 a.m. in a crowded strait.