Thesis - Comparison of different methods for individual salmon recognition

Company Description

We have been electrifying industries, powering homes and transforming life through innovation and collaboration for more than 100 years. Empowered by our goal, we are determined to enable fossil free living within one generation. To succeed we, as one of Europe's leading energy companies, must be fossil-free ourselves.

But that’s not enough. That’s why we are looking beyond our own industry, to see where we can really make a difference.

Join us in our journey towards a fossil-free life within one generation.

Want to learn more about what we do? Visit www.vattenfall.com.

Since Vattenfall is a part of the Swedish critical infrastructure, many of our services are security classed. If this position is security classed the final candidates might be subjected to a security vetting process, according to the Swedish legislation.

Job Description

Power Climate Smarter living - that’s our purpose. We are looking for the brightest students to make this happen. Together you will contribute to a fossil free future. You will have a unique position to contribute to our purpose. All we ask from you is to use your Superpowers and share your energy to help us achieve our goal. 

About the thesis

Department information

Research & Development is a department of about 130 employees at Vattenfall, working with laboratory tests, prototypes and new technologies to drive innovation and development in the company. You will be working in Asset Analytics, a team that uses AI and Data Science to analyse, monitor and improve Vattenfall’s assets. AI utilization for biodiversity monitoring is one of the main focuses of the team.

Thesis subject

The project is to be performed and supervised by SLU in Umeå.

Current technology of tracking individual fish in the wild relies on catching, handling and tagging fish to obtain positional data on their movement. This is costly, requires trained personnel and may cause stress and injury to the fish. Indeed, if the goal is to obtain high resolution data on fish movement over space and time, attaching active transmitters to the animal might be the only option. However, many research questions and management goals related to fish passage efficiency can be answered with just a few data points on each individual fish. Traditionally PIT technology (passive integrated transponder) have been used extensively for this purpose. It is a cost effective alternative of fish tracking but still requires the fish to be caught, tagged and handled. As an alternative, this project aims to use underwater camera traps to capture the spot patterns on migrating salmonids and use this for identification. The spots on salmonids are a unique signature for each individual and reported to be consistent over time (Stien et al. 2017), see Figure 1. Hence, it is possible to connect individual fish at different times and sites during their spawning migration, which would be an important tool when evaluating fish passage success and migration time for many individuals with a minimum effort.

The thesis project will be part of an ongoing project where Vattenfall has developed a system that uses image analysis to count and classify fish on their sort, origin, health status and sex. The next step in the project is to recognize individual fish and will be a collaboration between SLU and Vattenfall.

You will create an algorithm that can recognize individual fish based on these dot patterns. You will provide expertise around fish and their behavior, while Vattenfall provides knowledge around computer science and image recognition.

Goal

You will have access to two datasets with videos collected at different sites in Umeälven, one collected at the river mouth and the other from a fish ladder at the hydropower plant Stornorrfors. The goal of the project is to match the fish seen in the camera trap with the fish passing through the ladder. The proportion of successful fish passages and their migration time from the river mouth to passing the dam will provide valuable insight when evaluating measures to facilitate fish passage at the Stornorrfors hydropower plant.

Qualifications

We are looking for a student to join our journey towards a fossil free future. You are soon to be graduated from your academic studies. You also identify yourself with our principles: active, open, positive and safe.

Education - Conservation and Management of Fish and Wildlife Biology, or similar  

Scope - 30hp

Language - Fluent in English

Additional Information

• Start of assignment: 2024-01-22 or by agreement
• Location: Umeå or by agreement

• Application – a combined document with your CV and cover letter, and grades