Developed under real-world conditions

The specially developed, AI-based Dallmeier video analytics app, Sedor AI Ship Detection, detects large ships and small boats on open water surfaces – even under naturally difficult environmental conditions such as waves and water reflections. The core of the solution is a specially trained neural network that enables powerful AI object recognition of ships and boats. The solution was developed in cooperation with a port customer: Through a multitude of images and videos of real conditions, the neural network was able to consider and learn the specific requirements "at sea". Thus, the software can deliver reliable results by analysing the defined monitoring area in an adjustable cycle. If a ship or boat enters this area, the software issues a message. The false alarm rate is reduced by up to 99 percent compared to conventional analysis systems.

Hazard prevention in real time

"As a refinery with a tanker port, we are at risk of small watercraft entering the transhipment area without permission. The Dallmeier solution notifies us when such watercraft pass our port basin entrance," says Bernd Birnbacher, Head of Refinery Security at Holborn Europa Raffinerie. "This allows us to react immediately to a potential danger and once again significantly increase the level of security towards the water side. The Dallmeier solution also helps us to meet the requirements of the ISPS Code."

"We developed the video analysis app so that companies located by the sea can further improve their security and protection on the water side. However, application scenarios to improve business processes are also possible," says André Haack, Managing Director of Dallmeier Systems GmbH in Hamburg, which was entrusted with the implementation of the project. "For example, a container terminal operator can optimise its workflow by receiving automatic notifications when watercraft enter its port facility. Commercial port operators can also think of a variety of possible applications."