Forest damage is on the rise due to climate change, even in Hesse, one of Germany’s most forested states. The joint research project “SENSURE Wald,” led by the Northwest German Forest Research Institute (NW-FVA), the state-run HessenForst, and Hochschule RheinMain – University of Applied Sciences and Arts, has been developing AI models since this year that improve the spatial resolution of satellite data, thereby enabling changes in the forest to be detected more accurately and quickly. This should enable forest condition analyses to be conducted quickly and measures—such as reforestation or timber harvesting—to be promptly adapted to current conditions.
“With our new super-resolution method, we aim to locate even small areas of forest damage in Hesse’s forests in the future. Using machine learning methods—particularly AI—we generate high-resolution, more detailed images from satellite data provided by the EU’s Copernicus program. In addition, we are developing tailored AI models that enable the automated and efficient mapping of damaged forest areas with high accuracy and timeliness—without the previously time-consuming manual preparation,” explains Prof. Dr. Ulrich Schwanecke, a professor of Computer Science at HSRM.
Fast, High-Quality Information for Forestry Operations and Government Agencies
The planned models support sustainable, multifunctional forest management under the conditions of climate change. “We want to create digital tools for forestry operations and government agencies that quickly and reliably provide high-quality information about changes in the forest,” says Dr. Mengistu Abiy, head of the Forestry Geoinformation Division at the state agency HessenForst. A key element of this is the establishment of largely automated processes, as well as the development of a system that examines the derived forest damage data for plausibility. “All software components developed as part of the project—including the associated documentation—are to be published as open-source products and thereby made accessible to a broad range of users,” says Florian Franz of the NW-FVA. To make the developed applications usable in practical experience—for example, in forestry operations and government agencies—collaboration with public-sector IT companies is planned.
About the SENSURE Wald Research Project
The research project “Sentinel-2 Super-Resolution with Artificial Intelligence Methods for Remote-Sensing-Based Mapping of Forest Damage Areas” (SENSURE Wald; grant number 2224NR028B) is funded by the Federal Ministry of Agriculture, Food, and Home (BMLEH) as part of the “Sustainable Renewable Resources” funding program through the Agency for Renewable Resources (FNR) until September 2028. The total funding amounts to 633,000 euros.