Pubblicazioni Scientifiche

Filtri di ricerca 2 risultati
Reset
Pubblicazioni per anno
Benchmarking tree species classification from proximally sensed laser scanning data: Introducing the FOR-species20K dataset
Puliti , Stefano , Lines , Emily R. , Müllerová , Jana , Frey , Julian , Schindler , Zoe , Straker , Adrian , Allen , Matthew J. , Winiwarter , Lukas , Rehush , Nataliia , Hristova , Hristina S. , Murray , Brent A. , Calders , Kim , Coops , Nicholas C. , Höfle , Bernhard , Irwin , Liam A.K. , Junttila , Samuli , Kruček , Martin , Krok , G. , Král , Kamil , Levick , Shaun R. , Lück , Linda , Missarov , Azim , Mokroš , M. , Owen , Harry Jon Foord , Stereńczak , Krzysztof Jan , Pitkänen , Timo P. , Puletti , Nicola , Saarinen , Ninni , Hopkinson , Chris Dennis , Terryn , Louise , Torresan , C. , Tomelleri , Enrico , Weiser , Hannah , Astrup , Rasmus
lidar remote sensing biodiversity deep learning point cloud classification single-tree inventory
Mostra abstract
Proximally sensed laser scanning presents new opportunities for automated forest ecosystem data capture. However, a gap remains in deriving ecologically pertinent information, such as tree species, without additional ground data. Artificial intelligence approaches, particularly deep learning (DL), have shown promise towards automation. Progress has been limited by the lack of large, diverse, and, most importantly, openly available labelled single-tree point cloud datasets. This has hindered both (1) the robustness of the DL models across varying data types (platforms and sensors) and (2) the ability to effectively track progress, thereby slowing the convergence towards best practice for species classification. To address the above limitations, we compiled the FOR-species20K benchmark dataset, consisting of individual tree point clouds captured using proximally sensed laser scanning data from terrestrial (TLS), mobile (MLS) and drone laser scanning (ULS). Compiled collaboratively, the dataset includes data collected in forests mainly across Europe, covering Mediterranean, temperate and boreal biogeographic regions. It includes scattered tree data from other continents, totaling over 20,000 trees of 33 species and covering a wide range of tree sizes and forms. Alongside the release of FOR-species20K, we benchmarked seven leading DL models for individual tree species classification, including both point cloud (PointNet++, MinkNet, MLP-Mixer, DGCNNs) and multi-view 2D-based methods (SimpleView, DetailView, YOLOv5). 2D Image-based models had, on average, higher overall accuracy (0.77) than 3D point cloud-based models (0.72). Notably, the performance was consistently >0.8 across scanning platforms and sensors, offering versatility in deployment. The top-scoring model, DetailView, demonstrated robustness to training data imbalances and effectively generalized across tree sizes. The FOR-species20K dataset represents an important asset for developing and benchmarking DL models for individual tree species classification using proximally sensed laser scanning data. As such, it serves as a crucial foundation for future efforts to classify accurately and map tree species at various scales using laser scanning technology, as it provides the complete code base, dataset, and an initial baseline representative of the current state-of-the-art of point cloud tree species classification methods. © 2025 The Author(s). Methods in Ecology and Evolution published by John Wiley & Sons Ltd on behalf of British Ecological Society.
🔗 Vai alla pubblicazione
2025
Rivista
Methods in Ecology and Evolution
Dettagli
Vol. 16 (4) pp. 801 - 818
DOI
Towards an effective in-situ biodiversity assessment in European forests
Burrascano , Sabina , Chojnacki , Lucas , Balducci , Lorenzo , Chianucci , Francesco , Haeler , Elena , Kepfer-Rojas , Sebastian , Paillet , Yoan , de Andrade , Rafael Barreto , Boch , Steffen , de Smedt , Pallieter , Fischer , Markus , Mijangos , Itziar Garcia , Heilmann-Clausen , Jacob , Hofmeister , Jeňýk , Hošek , Jan , Kozák , Daniel , Kutszegi , Gergely , Lachat , Thibault , Mikoláš , Martin , Samu , Ferenc , Ravera , Sonia , Schall , Peter , Sitzia , Tommaso , Svoboda , Miroslav , Trentanovi , Giovanni , Ujházyová , Mariana , Vandekerkhove , Kris , Tinya , Flóra , Ódor , Péter
forest biodiversity vascular plants birds epiphytic bryophytes epiphytic lichens monitoring network multivariate standard error rarefaction curves saproxylic beetles wood-inhabiting fungi
Mostra abstract
Assessing multi-taxon biodiversity is crucial to understand forests’ response to environmental changes and to inform management strategies. In Europe, forest biodiversity monitoring is still scattered and heterogeneous, although a long-term monitoring network has long been advocated. Given the monitoring aims reported in various EU policies, this network should be accurately designed also through the estimation of its sampling effort, here intended as the number of sampling plots and sites. We used a novel database of forest multi-taxon biodiversity for a pilot study to: estimate the minimum sampling effort needed to: assess variation in species richness and composition; compare these estimates with the efforts invested in the pilot database; discuss estimates’ differences across taxonomic groups and forest categories. We focused on six taxonomic groups (vascular plants, birds, epiphytic lichens and bryophytes, wood-inhabiting fungi and saproxylic beetles) across six forest categories. Based on 6,165 plots at 2,084 different locations across Europe, we benchmarked the effort to achieve: a complete species richness estimate through interpolation/extrapolation curves, and a precise evaluation of species composition variation through multivariate standard error. Our estimates differed widely, especially among taxonomic groups. For species richness, estimates range from 3 to 147 plots per site across 3 to 29 sites per forest category, with birds and epiphytic bryophytes requiring the least effort. For species composition, estimates range from 5 to over 25 plots per site across 5 to 20 sites per forest category, with saproxylic beetles, vascular plants, and fungi displaying the highest estimates. The taxonomic groups requiring an effort comparable to existing data were the least diverse, all the others need greater efforts, either for species richness (e.g., saproxylic beetles), or species composition (e.g., vascular plants), or both (e.g., wood-inhabiting fungi). An effective monitoring network of European forests’ biodiversity should thoroughly account for these benchmarks and for their taxon-dependency. © 2025
🔗 Vai alla pubblicazione
2025
Rivista
Basic and Applied Ecology
Dettagli
Vol. 84 pp. 121 - 132
DOI