Unsupervised classification of LiDAR point clouds in tropical forests

Recent technological advancements in LiDAR scanning allows us to study the tropical forest structure in unprecedented detail. Methods proposed to date to analyze LiDAR point clouds are mainly supervised, and hence semi-automated. In this PhD, unsupervised machine learning strategies will be developed, extracting leaf and woody components from LiDAR point clouds. This will be a big step forward for operational applications.