Identify cold-water coral patterns from ROV video
MSc student project: How do cold-water coral distributions change on a carbonate mound?
Carbonate mounds are 10 to 100 m high structures in the deep sea, some of which are millions of years old. Carbonate mounds are the habitat of thriving cold-water coral reefs. These corals build a carbonate skeleton with which they trap sediment. This skeleton forms the basis of these huge mounds, so the mounds are effectively built by corals. Corals are limited by the amount of organic matter that reaches the seafloor.
There are all sorts of mechanisms that influence how organic matter is transported from the ocean surface to a carbonate mound. The amount of organic matter that reaches different places on a mound will affect the abundance at which the cold-water corals grow on the mound. It will likely also influence the pattern at which the corals grow. We have video-material from several transects with an ROV (Remotely Operated Vehicle) of a carbonate mound (see figure). We also have machine learning techniques available with which we can distinguish the different features in an image, such as ‘live corals’, ‘bare ground’, ‘dead corals’ etc.
We are looking for a motivated master student who wants to reveal hidden patterns in the images, by applying machine learning techniques on the ROV videos. Would you like to dive into the enigmatic world of the deep sea and help us understand cold-water coral growth? If you are up for a modelling challenge you can even help improve our machine learning scripts. Interest and affinity with programming are a big plus.