SFI SmartForest · Centre for Research-based Innovation · 2020–2029
SFI SmartForest (2020–2029) is a Centre for Research-based Innovation that digitalizes the Norwegian forest sector to make it more efficient, precise, and sustainable. NMBU plays a central role in the project, with NMBU’s part being led by Terje Gobakken. NMBU is responsible for the work packages on forest information (WP1 led by Erik Næsset) and environmental information (WP4 led by Hans Ole Ørka). The centre links emerging enabling technologies, such as AI and remote sensing, with forest mapping, silviculture, forest operations, and environmental information. This aims to create a world-leading R&D environment developing innovations for the entire value chain. The goal is to increase value creation, strengthen international competitiveness, and support the green shift by establishing a Norwegian forest-tech sector. At the same time, the initiative will improve the recruitment of young researchers and professionals to the industry.
Fig. 1. Work package structure.
See the project web pages for more info.
Project in NVA
- Project
- SFI SmartForest: Bringing Industry 4.0 to the Norwegian forest sector
- Period
- 2020–2028
- Funding
- Research Council of Norway (RCN) (309671)
- Coordinator
- Norwegian Institute of Bioeconomy Research
See the project and all its publications in NVA →
Results (18)
2026
- (2026). Assessing airborne laser scanning and aerial photogrammetry for convolutional neural network-based semantic segmentation of forest stands. Forestry (London), 99(5). DOI: 10.1093/forestry/cpag088
- (2026). FOR-age: Benchmarking individual tree age estimation using 3D deep learning on dense laser scanning data. Remote Sensing of Environment, 342. DOI: 10.1016/j.rse.2026.115462
- (2026). Enhancing tree species classification: insights from YOLOv8 and explainable AI applied to TLS point cloud projections. Forestry (London), 99(4). DOI: 10.1093/forestry/cpag059
- (2026). Deep learning analysis of forest machine operations for productivity analysis using 3D video classification, a pilot study. International Journal of Forest Engineering. DOI: 10.1080/14942119.2026.2665580
- (2026). Mapping forest road surface deterioration: detection, localization, and assessing maintenance features with the near-field sensor platform RoadSens. International Journal of Forest Engineering, 37(3), 457–470. DOI: 10.1080/14942119.2026.2665577
- (2026). ForestFormer3D: A Unified Framework for End-to-End Segmentation of Forest LiDAR 3D Point Clouds. IEEE International Conference on Computer Vision (ICCV), 24717–24727. DOI: 10.1109/iccv51701.2025.02292
- (2026). Toward automation in forest road classification: opportunities and limitations of geometry extraction from a mobile sensor platform. International Journal of Forest Engineering, 37(3), 443–456. DOI: 10.1080/14942119.2026.2656962
- (2026). Predicting wood assortment volumes from stand attributes, topography, and airborne laser scanning data using accurately positioned harvester data as a reference. Forestry (London), 99(2). DOI: 10.1093/forestry/cpag012
- (2026). Multispectral airborne laser scanning for tree species classification: A benchmark of machine learning and deep learning algorithms. ISPRS Journal of Photogrammetry and Remote Sensing (P&RS), 233, 278–309. DOI: 10.1016/j.isprsjprs.2026.01.031
- (2026). Forest road attributes and network demands characterized by changing climatic conditions in Norway. Norges miljø- og biovitenskapelige universitet (NMBU).
2025
- (2025). A precise deep learning approach for timber traceability along the forest value chain. International Journal of Forest Engineering. DOI: 10.1080/14942119.2025.2572938
- (2025). Impact of point cloud matching on precision and accuracy in area-based forest inventories. Silva Fennica, 59(3). DOI: 10.14214/sf.25002
- (2025). Assessing the effects of uncertainty in remote sensing-based predictions of tree species composition and site index on net present value calculations. Forestry (London), 1–14. DOI: 10.1093/forestry/cpaf057
- (2025). Site index determination in young Norway spruce forests using branch whorls detected from mobile laser scanning data by deep learning. Forestry (London). DOI: 10.1093/forestry/cpaf059
- (2025). Multi-objective value of information assessment using stochastic programming: addressing uncertainty in site index determination. Forestry (London), 9(1). DOI: 10.1093/forestry/cpaf021
- (2025). Towards enhancing field-based vegetation monitoring: A deep learning approach for species coverage estimation from ground-level imagery. Methods in Ecology and Evolution, 16(5), 949–957. DOI: 10.1111/2041-210x.70024
- (2025). Effects of sample tree selection and calculation methods on the accuracy of field plot values in area-based forest inventories. Silva Fennica, 59(2). DOI: 10.14214/sf.25003
- (2025). Semantic segmentation of forest stands using deep learning. arXiv, 60(1). DOI: 10.48550/arxiv.2504.02471
Publications linked to this project in NVA, including partners' work. Group members are in bold. Talks, posters and media appearances are listed in NVA.




