作者: András Zlinszky , Balázs Deák , Adam Kania , Anke Schroiff , Norbert Pfeifer
DOI: 10.5194/ISPRS-ARCHIVES-XLI-B8-1293-2016
关键词: Geography 、 Random forest 、 Habitat conservation 、 Vegetation 、 European union 、 Remote sensing 、 Lidar 、 Conservation status 、 Biodiversity 、 Environmental resource management 、 Natura 2000
摘要: Biodiversity is an ecological concept, which essentially involves a complex sum of several indicators. One widely accepted such set indicators prescribed for habitat conservation status assessment within Natura 2000, continental-scale programme the European Union. Essential Variables are designed to be relevant biodiversity and suitable global-scale operational monitoring. Here we revisit study 2000 mapping via airbone LIDAR that develops individual remote sensing-derived proxies every parameter required by manual, from perspective developing regional-scale Variables. Based on leaf-on leaf-off point clouds (10 pt/m2) collected in alkali grassland area, data products were calculated at 0.5 ×0.5 m resolution. These represent various aspects radiometric geometric texture. A Random Forest machine learning classifier was developed create fuzzy vegetation maps classes interest based these products. In next step, either classification results or selected as variables, fine-tuned field references. showed adequate performance summarized deliver with 80% overall accuracy compared This draws attention potential also holds implications mapping. (i) use sensor together habitat-level classification, (ii) utility seasonal data, including non-seasonal variables canopy structure, (iii) mapping-derived class probabilities species presence absence.