Our expertise in analysing and processing spatial data opens up innovative possibilities for accurately recording and monitoring forests and trees. By using modern machine learning techniques, neural networks and automated processes, we offer customised solutions ranging from individual tree detection to the close monitoring of forest damage and afforestation areas.
Single tree detection
Updating and expanding tree cadastres is a crucial step for modern green space management and forestry. With our machine learning methods, we not only identify the position of a tree, but also its crown diameter, tree height and other parameters. By combining a wide range of data sources (orthophotos, point clouds, surface models), we are able to capture significantly more trees than conventional cadastres. We achieve excellent results, particularly with free-standing trees.
Forest monitoring
For certification of sustainable forestry and to comply with the EU regulation on deforestation-free supply chains (EUDR), we are evaluating the possibilities of remote sensing for large-scale monitoring of forest areas. These innovative approaches are intended to supplement the previous on-site inspections and enable automated processes in the long term to prevent deforestation in global supply chains.
In an ongoing project in collaboration with the Irish Forestry Service, we are automatically detecting clear-cutting in forest areas throughout Ireland. Due to the challenges posed by cloud cover, which currently only allows for annual analyses instead of quarterly ones, we are continuously working to improve our methods, for example by using Sentinel-1 data.
Frühwarnsysteme für Waldschäden
We are working on the innovative potential of sun-induced fluorescence (SIF). Plants that carry out photosynthesis release excess energy in the form of fluorescence, which can be detected by new satellite missions such as FLEX or EnMAP. This can be used to identify damage to vegetation much earlier than with conventional spectral indices. An AI-based method that enables the early detection of declining plant health would also be conceivable.
Reforestation potential
On the web portals HERMOSA and ESMERALDA, which mundialis helped to develop, potential areas for reforestation and renaturation can be identified and monitored globally. In addition, these portals offer the opportunity to link local projects, global initiatives and donors. In cooperation with international partners, we thus create the basis for sustainable reforestation projects worldwide.