Subsea image analysis: A first national consortium assembled!

(post written by Matthias Obst, first published at Zooniverse)

In early March 2023 we assembled for the first time a national consortium around subsea image analysis for marine biology research and monitoring. The meeting brought together major national data providers such as the Swedish Meteorological and Hydrological Institute (SMHI), the Geological Survey of Sweden (SGU), Swedish National Dataservice (SND), the Swedish Agency for Marine and Water Management (SwAM), and the Swedish Biodiversity Data Infrastructure (SBDI). We also had several partner companies and NGO’s participating who actively work in the field of marine environmental monitoring (Combine, BeGeospatial, SeAnalytics, Medins, Wildlife.ai) as well as academic institutes (Lund University, University of Gothenburg).

The objective of the meeting was to explain the current state of the KSO image analysis system, present scientific use cases, and discuss plans for further development and integration with existing research infrastructures.

The workshop was a great success! We presented and discussed tangible use cases including applications for habitat mapping, community composition analysis, invasive species impact assessments, fish stocks estimations in marine protected areas, and environmental impact assessments in relation to wind park construction. Based on the agreements we achieved in the workshop we will now work towards user support as well as consolidation and documentation of the infrastructure services for use in ecological research and biological monitoring in Sweden.

More about subsea image analysis

For all of you who like to use the system or want to know more about it, just contact us or find more information and updates here:

Information page
Main repo
Jupyter notebooks (try them by pressing the “Open in Colab” button)
Current citizen science challenge

The development of the system is currently financed by the FORMAS project PLAN SUBSIM, the VINNOVA project Ocean Data Factory, as well as the VR infrastructure program Swedish Biodiversity Data Infrastructure (SBDI), and the European Digital Twin of the Ocean program (DTO).

Use cases in PLAN-SUBSIM

Cameras on moving robots: Sea pens and many other speceis identified in videos obtained by ROVs (Remotely Operated Vehicles)
Cameras in stationary observatories: Round gobies and many other species of fish observed using baited remote underwater video (BRUV). Photo: Leon Green
Bottom landing drop cameras: Mussel beds and other habitats identified by segmentation of images.