One focus of the Methods and Processing Group is the processing and analysis of image data from optical cameras for the exploration of maritime regions to detect and assess the situation above water.

Within your master thesis, you will investigate the applicability of synthetic scenes in comparison to real scenes for ship detection in maritime infrastructures. Using an existing realistic 3D model of a maritime infrastructure, as well as vessel models, you will work on the generation of synthetic images with photo-realistic appearance. Through a cross comparison of results obtained using synthetic/real images with computer vision and deep learning algorithms, the key objective will be an investigation of the impact of data augmentation using synthetic images to improve real da-tasets in tasks of computer vision for maritime security applications.

The main tasks will be:

• Set-up the provided 3D model of a maritime infrastructure and integration of 3D ship models within a 3D content creation tool
• Generation of synthetic images with the same appearance as a real ship dataset
• Cross comparison study of the real and synthetic dataset for ship detection or segmenta-tion using existing computer vision and deep learning algorithms

The scientific results of your master thesis will have the possibility to be followed by a publication in a conference or journal.