Assessing the potential of sentinel-1 and sentinel-2 satellite imagery for shoreline classification in support of oil spill preparedness and response

Yulmetova, Maria (2021) Assessing the potential of sentinel-1 and sentinel-2 satellite imagery for shoreline classification in support of oil spill preparedness and response. Masters thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (6MB)

Abstract

Coastal zones are critical ecosystems that provide important habitat for marine animals, fish, shellfish, birds, and many other species. However, there is a risk of mineral oil impacting in these areas due to human activities offshore. Shoreline classification is the first step to establishing response contingency plans in case of an oil spill. This study estimates the potential of using open-access, high-resolution Sentinel-1 and Sentinel-2 imagery for the mapping of shoreline types in support of oil spill preparedness and response activities. The two classification maps, depicting shoreline and coastal land cover, were produced using an advanced object-based Random Forest (RF) algorithm. Various features extracted from multi-source data, including spectral, texture, ratio, polarimetric features, and digital elevation model (DEM), were analyzed to identify the most valuable features for discrimination between different shoreline types. Multiple classification scenarios with the most useful features were then assessed and compared to find the best classification model. The developed algorithm achieved accuracies of 87.10% and 84.75% of coastal land cover and shoreline maps. These results demonstrated the high potential of using freely available Sentinel-1 and -2 satellite data for coastal mapping.

Actions (login required)

View Item