About this role

Across Europe, large-scale efforts are underway to restore degraded ecosystems and halt biodiversity loss. The EU Nature Restoration Regulation places new requirements on member states to monitor and report on the condition and restoration of various habitats, including wetlands, grasslands, forests, and other ecosystems. Wetlands are among the most valuable ecosystems for biodiversity, carbon storage, water regulation, and climate adaptation, but many have been degraded through historical land use changes. At the same time, rapid advances in Earth observation and machine learning (ML) create new opportunities for large-scale, cost-effective environmental monitoring. By combining satellite imagery, aerial data, and modern ML methods, it is increasingly possible to map and monitor habitat types and ecosystem conditions with high spatial detail.

In this master thesis , you will develop and train machine learning (ML) models for mapping Swedish wetlands, based on satellite and aerial imagery for Sweden. The work will include processing multimodal satellite images (e.g. high-resolution aerial orthophotos, DEM, Sentinel-1 and -2 data) and applying novel deep learning methods to perform fine-grained mapping of Swedish wetlands.

Techniques and models will include computer vision models (e.g., convolutional neural networks, vision transformers), as well as various data-efficiency techniques within ML. You will work in a multi-disciplinary team which includes both ML and biodiversity experts from RISE and the Swedish Environmental Protection Agency (Naturvårdsverket; project stakeholder).

The work requires students with excellent skills within ML, image processing, and preferably also remote sensing and/or GIS. As this is a master thesis project with a research organization, we will help you reach a high level of research excellence, and a successful project will ideally result in writing a joint research paper in addition to the master thesis.