Biophysical Characterization of Protected Areas Globally through Optimized Image Segmentation and Classification

摘要: Protected areas (PAs) need to be assessed systematically according biodiversity values and threats in order support decision-making processes. For this, PAs can characterized their species, ecosystems threats, but such information is often difficult access usually not comparable across regions. There are currently over 200,000 the world, assessing these ecological remains a huge challenge. However, linking remote sensing with modelling help overcome some limitations of conservation studies, as sampling bias inventories. The aim this paper introduce eHabitat+, habitat service supporting European Commission's Digital Observatory for Areas, specifically discuss component that stratifies into different functional types based on data. eHabitat+ uses an optimized procedure automatic image segmentation several environmental variables identify main biophysical gradients each PA. This allows systematic production key indicators compared globally. Results from few case studies illustrated show benefits open-source tool.