Datasets for robotic monitoring of EU habitats

To preserve biodiversity, the EU requires extensive habitat monitoring, as mandated by the EU Habitats Directive. Traditional methods rely on manual field surveys, which are labour-intensive and costly. The European Union NI project explores using robotics to assist this process, using quadrupedal robots and machine learning techniques that require large datasets. This paper presents four publicly available datasets collected by the quadrupedal robot ANYmal-C across four EU Natura 2000 habitats. These datasets include photos, videos of plant species, 3D point clouds, and robot locomotion data, offering valuable resources for research in species detection, habitat assessment, and robotics benchmarking.