Leaf-wood annotated tropical tree point clouds from terrestrial laser scanning

This dataset was used for the analysis of the following publication: Van den Broeck, W.A.J., Terryn, L., Chen, S., Cherlet, W., Cooper Z.T., Calders, K. (2025) Pointwise deep learning for leaf-wood segmentation of tropical tree point clouds from terrestrial laser scanning, ISPRS Journal of Photogrammetry and Remote Sensing, 227 (366-382), https://doi.org/10.1016/j.isprsjprs.2025.06.023 Any use of this dataset should cite the paper above (Creative Commons Attribution 4.0 International Public License). Contact: [e-mail] ---------------------------------------------------------------------------------------------- Dataset description: Dataset containing 148 manually leaf-wood annotated tropical tree point clouds. Point clouds were collected with a terrestrial laser scanner (RIEGL VZ-400) in northeastern Australia. The tree point clouds are given as .txt files, with the first three columns being the local x,y,z coordinates (z=height). The fourth column is the label (0 = leaf, 1 = wood). The trees are organised in a train, validation and test folder, to train and evaluate machine/deep learning models. Tee species information and more information on the plots can be found in the excel file 'additional_tree_and_plot_info.xlsx'. Code repository: https://github.com/qforestlab/leaf-wood-segmentation-with-deep-learning