@MASTERSTHESIS{pgi2020004,
	author = "D. Smith",
	supervisor = "M. Roper",
	title = "A Patch-Based Deep Learning Approach for Land-Classification of Sentinel-2 Satellite Imagery",
	school = "Department of Computer and Information Sciences, University of Strathclyde",
	year = "2019",
	abstract = "The application of machine learning to the remote sensing {\"\i}{\neg}eld has produced many exciting new advancements. One of these is in land-classi{\"\i}{\neg}cation: identifying the land-cover content present in satellite imagery. The use of deep neural networks has provided stateof-the-art accuracies for land-cover classi{\"\i}{\neg}cation. This project experiments with deep learning architectures and novel image handling to methods to improve the classi{\"\i}{\neg}cation accuracies achievable by Global Surface Intelligence, a geospatial analytics company. By implementing segmentation and patch-based training, we successfully show that by expanding their current methodology to consider spatial content, they can achieve increased classi{\"\i}{\neg}cation accuracies.",
}