Date of Award
4-2019
Degree Type
Research Paper
Degree Name
Master of Science in GIS for Development and Environment (GISDE)
Department
International Development, Community and Environment
Chief Instructor
Florencia Sangermano
Keywords
Sentinel-1, Synthetic Aperture Radar (SAR), Sentinel-2, Random Forest, Land Cover Classifcation
Abstract
This study evaluates Sentinel-1 and Sentinel-2 remotely sensed images for tropical land cover classification. The dual polarized Sentinel-1 VV and VH backscatter images and four 10-meter multispectral bands of Sentinel-2 were used to create six land cover classification images across two study areas along the border of the Bolivian Pando Department and the Brazilian state of Acre. Results indicate that Sentinel-2 multispectral bands possess a higher overall performance in delineating land cover types than the Sentinel-1 backscatter bands. Sentinel-1 backscatter bands delineated land cover types based on their surficial properties but did not facilitate the separation of similarly textured classes. The combination of Sentinel-1 and -2 resulted in higher accuracy for delineating land cover through increasing the accuracy in delineating the classes of secondary vegetation from exposed soil. While Sentinel-2 demonstrated the capability to consistently capture land cover in both case studies, there is potential for single date Sentinel-1 backscatter image to act as ancillary information in Sentinel-2 scenes affected by clouds or for increasing separability across classes of mixed multispectral qualities but distinct surficial roughness, such as bare ground versus sparsely vegetation areas.
Recommended Citation
Meneghini, Aaron, "An Evaluation of Sentinel-1 and Sentinel-2 for Land Cover Classification" (2019). Sustainability and Social Justice. 235.
https://commons.clarku.edu/idce_masters_papers/235
Included in
Geographic Information Sciences Commons, Physical and Environmental Geography Commons, Remote Sensing Commons
