Publication
A global AI Platform for live Detection of Marine Debris
Dennis Krummacker Maryam Arabshahi
In: Marlen Davis Klemens Mrogenda (Hrsg.). NaturschutzDigital – Künstliche Intelligenz im Naturschutz Forschung, Praxis und Leitplanken. Pages 40-41, Vol. BfN-Schriften (632.2022 -) (650), Deutschland / Bundesamt für Naturschutz, Bonn, 11/2023.
Abstract
Marine litter, be it large accumulations or sheer volume, is harmful for marine life as well as deteriorating the health of the human population. Besides better awareness and avoiding future pollution, also an important step is recognition and removal of already existing litter. The task of recognition requires two tasks: Capturing the status of the sea and processing this data to pinpoint the litter. AI proved to be very successful in identifying marine debris in radar data. To advance this towards live interaction, we propose a global Data-Analytics (DA) Platform. The user-level experience shall be that new measured data is automatically transmitted to the global database and that position and trajectory of debris can be viewed and predicted up-to-date. To enable this, the platform automatically integrates received data in time and utilizes AI to identify the litter. The sub-task of communication to and from the platform is very crucial, not only because a global interconnection is desired, but also certain aspects in a following concrete system design can be inspired by modern and future communication systems. Given that the application environment is expected to happen mostly distant to land, i.e. where no sophisticated communication infrastructure can be assumed, the bare physical interconnection is a concern. For that, satellite communication might become a relevant solution, which is also heavily discussed in the course of mobile communications, such as recent 6G research.