Development of Supervised and Unsupervised Computer Vision Methods to Quantify Salmon and Forage Fish Biodiversity from Underwater Video

Underwater video cameras are a convenient method to monitor aquatic life such as fish. However, these cameras can collect a lot of footage, which can be challenging for humans to process. This project will use computer vision to automatically sort through video taken within and near to seaweed farms to specify the clips that have fish in them. The computers will then help identify and count rare and common fish to describe how marine organisms use seaweed farms as habitat. This information helps companies like Cascadia Seaweed farm seaweed in better ways to promote marine biodiversity by making their farms better habitats for marine life.

Faculty Supervisor:

Alexandra Branzan Albu;Francis Juanes

Student:

Partner:

Cascadia Seaweed

Discipline:

Computer science

Sector:

Sustainability & the Environment; Artificial Intelligence; Aquaculture and Fishing

University:

University of Victoria

Program: