AI toepassingen in de scheepvaart.
Geschreven door: Chris Twigt, Competence Developer AI bij Pancompany.
Hij was aanwezig op een Data & Drinks event georganiseerd door Xomnia en schreef daar het volgende over.
Steering inland ships – Pim Dijt & David Woudenberg (Shipping Technology)
Shipping Technology has developed a lane assist system for inland ships. The system is modular and connects to all existing sensors and actuators on board of a ship. There are several challenges: Integration with existing sensors and actuators of infinite variety, ships composed for barges left right and center, influence of weather, lateral movement of the ship and varying inertia depending on the load. All in all an impressive system that is still being simplified by reducing the number of models used.
Measuring draught of passing ships with AI – Michael Dubbeldam (Technolution)
When ships enter The Netherlands at the port of Ijmuiden they must be checked for draught to prevent them from entering water that is too shallow. The solution consists of on-shore cameras that are controlled to track passing ships and detect- and interpret the water mark(s) on the ship and the water level at its hull to get to its draught. The solution uses a couple of IoT devices and a high quality orientable camera with infrared support. Challenges here are the reflections of the water on the hull, averaging out waves and the latency in the entire chain (receiving video frame -> mark detection -> pan/tilt camera command -> pan/tilt camera) which was over 2 seconds in total. The low amount of compute it uses is impressive. A truly elegant solution and inspiring presentation!
It is refreshing to see applications of AI on the edge amidst the tsunami of news on LLM’s and RAG applications we are currently experiencing in the market.