Beyond the Forest - Multi-Decadal Mapping of Trees In and Outside Forests Using Deep Learning

PhD defence by Rene Arum Rafael Lee

Assessment Committee

Associate Professor Andreas Westergaard-Nielsen, Department of Geosciences and Natural Resource Man, University of Copenhagen (Chairperson)

Agronomy Research Scientist Agnes Begue, French Agricultural Research Centre for International Development (CIRAD)

Senior Researcher Zander Venter, Norwegian Institute for Nature Research (NINA)

Supervisors

Professor Rasmus Fensholt
Associate Professor Martin Brandt

Department

Department of Geosciences and Natural Resource Management

Place

The defence is conducted as a hybrid defence.

To attend the defence in person:
Building: Institut for Geovidenskab og Naturforvaltning, Room: Festsalen,
Øster Voldgade 10, 1350 Copenhagen

To attend the defence online:
Please follow the link to attend the defence online: https://ucph-ku.zoom.us/j/66717566969
MeetingID, if relevant: 667 1756 6969

Email address to gain access to the thesis: rene.lee@ign.ku.dk.
You will either receive a copy of the thesis or be informed where you can read a physical copy.
Recipients of copies of the thesis are not allowed to share or distribute it due to copyright compliance.

Short description of the thesis

This thesis aims to improve the monitoring of trees in and outside of
forests over the past three decades, using historical satellite imagery and
deep learning. A new method is developed for bridging the gap between
modern high resolution tree maps and low resolution, long-term forest
maps, to uncover the dynamics of trees outside forests that have
previously gone unreported. A learned disturbance mapping framework
follows on as well as a national forest inventory constrained tree cover
map resulting from the fusion of 11 existing forest maps, to create a robust
forest cover monitoring system.