Digital data collection is an essential part of almost all underwater missions. In addition to explicit data collection as the purpose of the mission, for example in the inspection of underwater structures or the digital acquisition of the seafloor, very large amounts of data are also generated during mission control (especially for autonomous vehicles). These heterogeneous, distributed, and large-volume data must be easily combined and interpreted in the context of existing data. The project aims to make the sensor data, which has so far mostly been collected separately, available as fused context data in a framework as a basis for developments in the cluster. A key aspect is the efficient application of machine learning (ML) methods to complex data types such as 3D point clouds or stereo videos. A particular challenge is multi-sensor fusions (camera, acoustics, flow) in a physically demanding environment including narrowband connections between the individual systems.