Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale

Research output: Contribution to journalJournal articlepeer-review

Standard

Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale. / Li, Sizhuo; Brandt, Martin; Fensholt, Rasmus; Kariryaa, Ankit; Igel, Christian; Gieseke, Fabian; Nord-Larsen, Thomas; Oehmcke, Stefan; Carlsen, Ask Holm; Junttila, Samuli; Tong, Xiaoye; d’Aspremont, Alexandre; Ciais, Philippe.

In: PNAS Nexus, Vol. 2, No. 4, pgad076, 2023.

Research output: Contribution to journalJournal articlepeer-review

Harvard

Li, S, Brandt, M, Fensholt, R, Kariryaa, A, Igel, C, Gieseke, F, Nord-Larsen, T, Oehmcke, S, Carlsen, AH, Junttila, S, Tong, X, d’Aspremont, A & Ciais, P 2023, 'Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale', PNAS Nexus, vol. 2, no. 4, pgad076. https://doi.org/10.1093/pnasnexus/pgad076

APA

Li, S., Brandt, M., Fensholt, R., Kariryaa, A., Igel, C., Gieseke, F., Nord-Larsen, T., Oehmcke, S., Carlsen, A. H., Junttila, S., Tong, X., d’Aspremont, A., & Ciais, P. (2023). Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale. PNAS Nexus, 2(4), [pgad076]. https://doi.org/10.1093/pnasnexus/pgad076

Vancouver

Li S, Brandt M, Fensholt R, Kariryaa A, Igel C, Gieseke F et al. Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale. PNAS Nexus. 2023;2(4). pgad076. https://doi.org/10.1093/pnasnexus/pgad076

Author

Li, Sizhuo ; Brandt, Martin ; Fensholt, Rasmus ; Kariryaa, Ankit ; Igel, Christian ; Gieseke, Fabian ; Nord-Larsen, Thomas ; Oehmcke, Stefan ; Carlsen, Ask Holm ; Junttila, Samuli ; Tong, Xiaoye ; d’Aspremont, Alexandre ; Ciais, Philippe. / Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale. In: PNAS Nexus. 2023 ; Vol. 2, No. 4.

Bibtex

@article{ccf1c1e94f3047279db851320ce81a21,
title = "Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale",
abstract = "Sustainable tree resource management is the key to mitigating climate warming, fostering a green economy, and protecting valuable habitats. Detailed knowledge about tree resources is a prerequisite for such management but is conventionally based on plot-scale data, which often neglects trees outside forests. Here, we present a deep learning-based framework that provides location, crown area, and height for individual overstory trees from aerial images at country scale. We apply the framework on data covering Denmark and show that large trees (stem diameter >10 cm) can be identified with a low bias (12.5%) and that trees outside forests contribute to 30% of the total tree cover, which is typically unrecognized in national inventories. The bias is high (46.6%) when our results are evaluated against all trees taller than 1.3 m, which involve undetectable small or understory trees. Furthermore, we demonstrate that only marginal effort is needed to transfer our framework to data from Finland, despite markedly dissimilar data sources. Our work lays the foundation for digitalized national databases, where large trees are spatially traceable and manageable.",
author = "Sizhuo Li and Martin Brandt and Rasmus Fensholt and Ankit Kariryaa and Christian Igel and Fabian Gieseke and Thomas Nord-Larsen and Stefan Oehmcke and Carlsen, {Ask Holm} and Samuli Junttila and Xiaoye Tong and Alexandre d{\textquoteright}Aspremont and Philippe Ciais",
year = "2023",
doi = "10.1093/pnasnexus/pgad076",
language = "English",
volume = "2",
journal = "PNAS Nexus",
issn = "2752-6542",
publisher = "National Academy of Sciences",
number = "4",

}

RIS

TY - JOUR

T1 - Deep learning enables image-based tree counting, crown segmentation and height prediction at national scale

AU - Li, Sizhuo

AU - Brandt, Martin

AU - Fensholt, Rasmus

AU - Kariryaa, Ankit

AU - Igel, Christian

AU - Gieseke, Fabian

AU - Nord-Larsen, Thomas

AU - Oehmcke, Stefan

AU - Carlsen, Ask Holm

AU - Junttila, Samuli

AU - Tong, Xiaoye

AU - d’Aspremont, Alexandre

AU - Ciais, Philippe

PY - 2023

Y1 - 2023

N2 - Sustainable tree resource management is the key to mitigating climate warming, fostering a green economy, and protecting valuable habitats. Detailed knowledge about tree resources is a prerequisite for such management but is conventionally based on plot-scale data, which often neglects trees outside forests. Here, we present a deep learning-based framework that provides location, crown area, and height for individual overstory trees from aerial images at country scale. We apply the framework on data covering Denmark and show that large trees (stem diameter >10 cm) can be identified with a low bias (12.5%) and that trees outside forests contribute to 30% of the total tree cover, which is typically unrecognized in national inventories. The bias is high (46.6%) when our results are evaluated against all trees taller than 1.3 m, which involve undetectable small or understory trees. Furthermore, we demonstrate that only marginal effort is needed to transfer our framework to data from Finland, despite markedly dissimilar data sources. Our work lays the foundation for digitalized national databases, where large trees are spatially traceable and manageable.

AB - Sustainable tree resource management is the key to mitigating climate warming, fostering a green economy, and protecting valuable habitats. Detailed knowledge about tree resources is a prerequisite for such management but is conventionally based on plot-scale data, which often neglects trees outside forests. Here, we present a deep learning-based framework that provides location, crown area, and height for individual overstory trees from aerial images at country scale. We apply the framework on data covering Denmark and show that large trees (stem diameter >10 cm) can be identified with a low bias (12.5%) and that trees outside forests contribute to 30% of the total tree cover, which is typically unrecognized in national inventories. The bias is high (46.6%) when our results are evaluated against all trees taller than 1.3 m, which involve undetectable small or understory trees. Furthermore, we demonstrate that only marginal effort is needed to transfer our framework to data from Finland, despite markedly dissimilar data sources. Our work lays the foundation for digitalized national databases, where large trees are spatially traceable and manageable.

U2 - 10.1093/pnasnexus/pgad076

DO - 10.1093/pnasnexus/pgad076

M3 - Journal article

C2 - 37065619

VL - 2

JO - PNAS Nexus

JF - PNAS Nexus

SN - 2752-6542

IS - 4

M1 - pgad076

ER -

ID: 339738474