Manual crown delineation of individual trees in 2 countries: Denmark and Finland

Dataset download link: https://sid.erda.dk/sharelink/eFt21tspNe

GitHub page: https://github.com/sizhuoli/TreeCountSegHeight

Publication: https://doi.org/10.1093/pnasnexus/pgad076

Dataset description

---------- Denmark data---------- :

More than 20k individual trees growing in different landscapes.

Annotation masks are paired with aerial photographs with a spatial resolution of 20cm.

Images contain RGB and near-infrared bands and were taken in summer 2018.

Image credits: Danish Agency for Data Supply and Infrastructure (https://sdfi.dk/)

---------- Finland data---------:

More than 4k individual tree crowns in random sampling regions

Annotation masks are paired with aerial photographs containing near-infrared, green, and blue bands with a spatial resolution of 50cm

Image credits: National Survey of Finland (https://www.maanmittauslaitos.fi/en/maps-and-spatial-data)

Tasks

Segmentation of individual tree crowns
Transfer learning/domain adaptation between datasets with different visual semantics, band compositions, forest species, etc.

Citation

@article{li2023deep,
  title={Deep learning enables image-based tree counting, crown segmentation, and height prediction at national scale},
  SHORTauthor={Li, Sizhuo and Brandt, Martin and Fensholt, Rasmus and Kariryaa, Ankit and Igel, Christian and Gieseke, Fabian and Nord-Larsen, Thomas and Oehmcke, Stefan and Carlsen, Ask Holm and Junttila, Samuli and others},
author={Li, Sizhuo and Brandt, Martin and Fensholt, Rasmus and Kariryaa, Ankit and Igel, Christian and Gieseke, Fabian and Nord-Larsen, Thomas and Oehmcke, Stefan and Carlsen, Ask Holm and Junttila, Samuli and Xiaoye Tong and Alexandre d’Aspremont and Philippe Ciais},
  journal={PNAS nexus},
  volume={2},
  number={4},
  year={2023},
  publisher={Oxford University Press}
}