Ocean colour mapping using remote sensing technology and an unsupervised machine learning algorithm

Parsons, Sarah and Amani, Meisam and Moghimi, Armin (2021) Ocean colour mapping using remote sensing technology and an unsupervised machine learning algorithm. Journal of Ocean Technology, 16 (3). pp. 102-115. ISSN 1718-3200

[English] PDF Available under License Creative Commons Attribution Non-commercial. Download (545kB)

Abstract

Satellites allow users to observe ocean colour in a way that is not possible from a ship or the shore. Ocean colour depends on interactions of incident light with particles or substances in the water. These light interactions cause the ocean to be a variety of shades depending on what the water is composed of and how these materials change the reflections of the light. The ocean colour fluctuation can be caused by different compositions, such as the biomass of phytoplankton or zooplankton, and can lead to a change in ocean colour, for example, from normal clear blue into a variety of shades of green. Satellites take measurements that can be used to calculate ocean colour and concentrations of materials in the ocean. This study focused on ocean colour mapping using satellite images captured from the Mediterranean Sea. The Iterative Self Organizing Data Analysis Techniques Algorithm (ISODATA) unsupervised machine learning (ML) algorithm was employed to determine ocean colour. The produced map is a basic way of displaying ocean colour and is easy for users of any skill level to produce. Finally, it was observed that having more information about phytoplankton and applying it to the algorithm could improve the results

Actions (login required)

View Item