Elevate

Environmental monitoring creates a foundation for understanding our natural environment and making informed policy and management decisions. In Sweden, monitoring of open landscapes and habitats is gathered under the NILS (National Inventories of Landscape in Sweden) program. One such habitat is the alpine areas in Northwestern Sweden. The NILS alpine inventory relies on a two-stage inventory design where wall-to-wall models based on satellite and field data predict which areas that should be visited by field personnel. Ensuring that the correct areas are visited is essential for a cost-effective inventory, underscoring the importance of accurate model predictions. New technical solutions could assist in creating even better models. 

In recent years UAVs (drones) have emerged as a mature and versatile tool for data collection. Paired with high-resolution cameras and advanced photogrammetry software, high-quality orthophotos and point clouds can be created. Such data sets can be used for a variety of tasks such as object detection and vegetation cover mapping. Drone imagery can also be used as a bridging solution between field inventories and relatively low-resolution satellite imagery, creating a “near ground truth” data set. Another emerging area of research is computer vision using DNNs (Deep Neural Networks). Such algorithms paired with drone and satellite data have been successfully used to classify vegetation cover in multiple contexts, however experiences of implementation into long-term monitoring programs remain limited.

ELEVATE addresses these challenges by integrating UAV imagery collected by NILS field personnel with satellite scenes, LiDAR data, and aerial imagery to develop robust DNN-based workflows for estimating vegetation cover, type, and change. The project started in July 2025 and is expected to finish mid-2028, with the goal to deliver a generalized and useable model for project stakeholders.