The key project for the Oceanic Machine Vision group is DEEP QUANTICAMS (2019-2024), funded by the German Research Foundation through the Emmy Noether Programme. The goal of the project is to enable measuring, mapping and monitoring with cameras in the deep sea by establishing new models, calibration and observation technologies for machine vision in the ocean, and to apply the technologies to ocean science problems.
OMV also participates in the EU project TechOceanS (2020-2024), where we work jointly with European partners on plankton observation, simulation and underwater light calibration to improve sensing the essential ocean variables.
The need to apply modern pattern recognition methods in the ocean sciences led to the foundation of a new graduate school for marine data scientists. OMV also collaborates in MarDATA, training PhD students, e.g. using quantitative or calibrated underwater imaging and automating visual observations.
The project ProBaNNt aims at AI for supporting the handling process of world war munitions found at the seafloor. A core part of this project lies in detailed 3D reconstruction of UXOs from super high resolution acoustic and optical data to assist identification, state estimation and planning of treatment.
In all projects we employ versatile computer vision methods, to tackle various challenges arising from the automated processing of underwater optical observations. Please pick some of the following Research Topics for a detailled impression: