Aquatic research often relies on image and video data to gain knowledge about our oceans, lakes and rivers. Current challenges in this research area include analyzing human-caused littering or oil spills. These investigations are based, for instance, on drone or satellite images of the water surface. Although intensive research is being conducted in this area, there is still a lack of uniform frameworks and platforms that can simplify the processing of image data in aquatic applications to make it more efficient and accessible.
This gap will be closed by the iMagine platform, which builds on the already existing European Open Science Cloud (EOSC) platform. The EU-funded project aims to build a platform that provides a set of AI-based image analysis services and a repository of image data provided by several research institutes. By additionally highlighting best practices, the aim is to provide easy access to AI-based image processing services so that potential users, independent of their knowledge of AI, can benefit from the applications.
This project receives funding from the European Unions horizon programme under grant agreement number 101058625.
Future users can utilize the platform in different ways, depending on their individual experience with AI applications. By providing pre-trained models, it is possible via a user-friendly API to use AI models for prediction without having to train them themselves. The provided models can also be re-trained to fit the individual input data. The framework also enables the development of custom Deep Learning models by providing resources for data storage as well as cloud resources for the deployment of the developed models.
Overall, the project aims to improve existing image analysis techniques through the use of AI, thereby enhancing the quality of image-based maritime research. To achieve this, 12 research institutes are working on eight use cases in the maritime environment, which will be integrated into the platform. The DFKI research group Marine Perception is part of this project and provides a system for the identification and quantification of floating and shore-washed plastic litter. For this use case, DFKI is taking the lead and supports the project with its expertise.
Partners
- Centro Euro-Mediterraneo sui Cambiamenti Climatici - CSIC (Spanish National Research Council) - EGI Foundation - European Multidisciplinary Seafloor and water column Observatory (EMSO-ERIC) - GRNET Consortium - Institut Francais de Recherche pour l'Exploitation de la Mer (Ifremer) - Foundation for Science and Technology (FCT) - Institute of Informatics, Slovak Academy of Sciences (IISAS) - Karlsruhe Institute of Technology (KIT) - Universite de Lorrainse (UL-LIEC) - Laboratory for Instrumentation and Experimental Particle Physics (LIP) - Marine Institute (MI) - Mariene Informatie Service MARIS BV (MARIS) - Istituto Nazionale di Oceanografia e di Geofisica Sperimentale (OGS) - Orbital Eos Sl (OrbitalEOS) - Sistema d’observació i predicció costaner de les Illes Balears (SOCIB) (SOCIB) - Sorbonne Universite (SU) - Turkish Academic Network and Information Centre (ULAKBIM) - Universita Degli Studi di Trento (UNITN) - Universitat Politecnica de Catalunya (UPC) - Vlaams Instituut voor de Zee (VLIZ) - Walton Institute
