Forest Monitoring in Guatemala Using Satellite Imagery and Deep Learning

Abstract

Forests cover 34% of Guatemala. The Guatemalan government have taken proficient actions in past decades to reduce deforestation and are looking toward new space technologies to improve forestry monitoring. This paper demonstrates the ability to automatically detect pixel-level changes in satellite images of forested areas that can be used to assist Guatemalan agencies, using satellite imagery from the Copernicus program and specially-developed deep learning algorithms for image segmentation.