Automatic analysis of the shape of trees from point clouds acquired in a forest

This thesis presents a pipeline for the automatic analysis of tree structure from 3D point clouds acquired in a forest environment. The first step involves the segmentation of individual trees using existing models. Once segmented, each tree is processed through a skeletonization phase to extract a simplified representation of its structure. Based on the resulting skeleton, branches are identified and characterized. Finally, a morphological fingerprint is generated for each tree, enabling further analysis and comparison.