Ocean color is an essential tool to understand the oceans that surrounds us. Traditional Ocean Color Sensors utilize multispectral data. Data from these sensors sample the electromagnetic spectrum at a much coarser resolution than newer hyperspectral sensors. We want to leverage the heritage from these legacy instruments to provide better estimations of the biomass using hyperspectral data. In this project thesis, the student should investigate how to use ensemble techniques to give both inference and uncertainty of the biomass.

This work will build on the data provided in [1]. A good starting point for inspiration is the IOCCG reports[2], especially [3]. A data source can be found here [4]. Here is some information on ensemble models [5]

For more information or a short discussion, contact sivert.bakken@ntnu.no

1. https://www.sciencedirect.com/science/article/pii/S003442571930166X?via%3Dihub
2. https://ioccg.org/what-we-do/ioccg-publications/ioccg-reports/
3. https://ioccg.org/wp-content/uploads/2015/10/ioccg-report-07.pdf
4. https://doi.pangaea.de/10.1594/PANGAEA.862886
5. https://towardsdatascience.com/ensemble-models-5a62d4f4cb0c