Robotic monitoring of European habitats: a labeled dataset for plant detection in Annex I habitats of Italy

Abstract

The present data descriptor presents a dataset designed for the detection of plant species in various habitats of the European Union. This dataset is based on images captured using multiple different hardware including quadrupedal robot ANYmal C, referring to ecologically important species to assess the presence and conservation status in Annex I habitats 2110, 2120, 6210*, 8110, 8120, and 9210*. Plant scientists and robotic engineers gathered the data in key Italian protected areas and labeled it using YOLOtxt format. Researchers in vegetation science, habitat monitoring, robotics, machine learning, and biodiversity conservation can access the dataset through Zenodo. The ultimate goal of this collaborative effort was to create a dataset that can be used to train artificial intelligence models to assess parameters that enable robotic habitat monitoring. The availability of this dataset may enhance future studies and conservation initiatives for Annex I habitats inside and outside the Natura 2000 network. The dataset and the methods used to obtain it are fully described, highlighting the significance of interdisciplinary cooperation in habitat monitoring.