Hopp til hovedinnholdet

Search

allprojectspublicationssubjectsemployeesserviceseventsnews
Back to projects

Division of Forest and Forest Resources

PathFinder - Towards an Integrated Consistent European LULUCF Monitoring and Policy Pathway Assessment Framework

Pathway assesment graphic
Photo:
Active Last updated: 31.01.2026
End: aug 2026
Start: sep 2022

More information

Official project website

Project participants

Johannes Schumacher Rasmus Astrup Stefano Puliti Clara Antón Fernandéz Ryan Bright Morgane Merlin Zsofia Koma
Partners

Albert-Ludwigs University Freiburg (ALU), Germany

National Institute of Geographic and Forest Information (IGN), France

Vrije Universiteit Amsterdam (VUA), Netherlands

Thünen Institute of Forest Ecosystems (TI), Germany

Croatian Forest Research Institute (CFRI)

Natural Resources Institute Finland (LUKE)

Federal Research and Training Center for Forests, Natural Hazards and Landscape (BFW), Austria

Slovenian Forestry Institute (GIS)

Czech Forest Management Institute (UHUL)

Technical Research Centre of Finland Ltd. (VTT)

Consejo Superior de Investigaciones Científicas (CSIC), Spain

Center for International Climate Research (CICERO), Norway

University of Göttingen (UGOE), Germany

University of Helsinki (UH), Finland

TreeMetrics (TM), Ireland

Eigen Vermogen van het Instituut voor Natuur- en 
Bosonderzoek (EVINBO), Belgium

European Landowners Organisation (ELO), Belgium

Institut Européen de la Forêt Cultivée (IEFC), France

Finnish Meteorological Institute (FMI)

 

Associated partners:

Swiss Federal Research Institute for Forests Snow and Landscape Research (WSL)

University of Bristol (UB), United Kingdom

Joint Research Center (JRC), Belgium

European Environmental Agency (EEA), Denmark

EuropeMap_partners_subcontractors_v2.png
Partners and supporters of PathFinder
Status Active
Start - end date 01.09.2022 - 31.08.2026
Project manager Johannes Breidenbach
Division Division of Forest and Forest Resources
Department National Forest Inventory
Total budget 57492664
Forests play an essential role in human well-being and health as well as provide a wide array of functions such as biodiversity protection, wood production, and climate change mitigation. To manage forests in a sustainable way requires efficient forest management strategies to obtain detailed and accurate information. The EU-funded PathFinder project aims to facilitate this process by going beyond current state-of-the art practices to make the most efficient use of field and remotely sensed data by creating high-resolution maps and accurately estimating forest attributes. All in all, forest monitoring will prove to be beneficial for decision making processes and policy formulation at the regional, national, and European levels.
Pathfinder final 1.png

Publications in the project

Authors

Johannes Breidenbach

Abstract

No abstract has been registered

Authors

Johannes Breidenbach

Abstract

No abstract has been registered

Authors

Johannes Breidenbach

Abstract

No abstract has been registered

Authors

Johannes Breidenbach

Abstract

Presentation

To document

Authors

Pauline Müller Stefano Puliti Johannes Breidenbach

Abstract

1. Field-based vegetation mapping is important for environmental assessments.Often, the area covered by a species is estimated visually within a reference frame.However, such assessments are prone to observer bias and a large variability. 2. We developed a deep learning pipeline relying on YOLOv8 models to segmentspecies and estimate the percentage cover (%) of Vaccinium myrtillus (blueberry)and Vaccinium vitis-idaea (lingonberry), two key understory species in borealforests. We used 138 nadir and downward-looking images of the forest floorcaptured in correspondence with 50 × 50 cm vegetation sub-plots assessedwithin National Forest Inventory (NFI) plots. First, we trained a bounding-boxframe detection model to crop the image to the same area assessed in the field.Second, we trained an instance segmentation model to classify species. Third,we flattened the class values into a semantic raster and estimated the species-specific cover by pixel counting. 3. We evaluated our method against an independent test set of 156 images andfound a root mean squared error (RMSE) of 8.82% for blueberry and 3.49% forlingonberry and no substantial systematic errors. An additional comparison withocular estimation by various field workers for the same plots showed that themodel estimates were within the range of estimates by field workers 8 out of 9times for blueberry and 7 out of 9 times for lingonberry. 4. The developed method shows promise in reducing observer bias and variabilityin vegetation surveys, thereby improving their consistency while significantlyreducing the time needed for species-specific coverage estimation. This isparticularly beneficial for repeated measurements and monitoring vegetationcover dynamics. However, as the method relies on RGB data, it is limited toestimating the percentage of visible species that are not obscured by others.Expanding the method to include a broader range of cover classes (e.g. grasses,rocks, logs) or species could automate the capture of crucial information

Show more publications
Other projects from National Forest Inventory