This dataset includes manually delineated individal tree crowns of urban trees in Berlin and Osnabrueck. These tree crowns were manually labelled based on official tree register data and are divided into seven different urban environment classes. The Mask R-CNN prediction results based on orthophoto imagery from August 2020 for Berlin and June 2023 for Osnabrueck were also included in the dataset. These, in combination with the ground truth labels, were used for tuning and validating the Mask R-CNN model. Overall, this dataset features aerial imagery in false colour representation, ground truth labels and the predicted tree crown delineations created by our model. This data can be used for further ITCD analysis or advanced urban tree monitoring. A publication based on this dataset is available at:https://doi.org/10.1109/JSTARS.2025.3626942 If you use this dataset or the corresponding publication, please cite them as follows: Dataset:Schütte, O. (2025). Individual Tree Crown Delineation Datasets of Berlin and Osnabrueck [Data set]. Zenodo. https://doi.org/10.5281/zenodo.15596557 Publication:O. Schütte, M. Lucas and B. Waske (2025). Individual Tree Crown Delineation in Heterogenous Urban Environments using High-Resolution Airborne Imagery. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing. vol. 18, pp. 27923-27936, 2025. https://doi.org/10.1109/JSTARS.2025.3626942 For more Information please read: README