Research at Lendspitz

Fotocredit: FH Kärnten
alt text
Management of Conservation Areas MSc

In summer and autumn 2021, two drone (UAS) flights were performed at the Natura 2000 site Lendspitz-Maiernigg, Klagenfurt. UAS allows cost-efficient habitat monitoring in protected areas over large spatial scales and offer opportunities for automated identification and classification of certain species or habitats.

Moreover, different vegetation indices derived from aerial images give hints on habitat conditions like water availability or growth-performance of the vegetation. Whilst this is a common approach with satellite data, the use of drone images for similar purposes is less common, though becoming more standard.

The BioMONITec team is testing such new technologies for their applicability in protected area management. For example, UAS-technology could give near-real-time information about habitat conditions or changes in abundance of habitat types and could also be used for detection of illegal activities like logging or poaching. These applications require highly automatised image classification algorithms, especially for usability over larger spatial areas.

For automatic or semi-automatic classification of aerial images, the BioMONITec -team is currently testing several different algorithms. Among them are fully automated classification approaches based on pixel values, but also semi-automated ‘random-forest’ approaches where an algorithm is trained with pre-classified data. Aerial images are combined with digital surface models in order to improve habitat classifications. In addition, different vegetation indices are tested regarding their usability to determine habitat conditions.