AerialWaste is a dataset for the discovery of illegal landfills. Illegal landfills from aerial images present a visual heterogeneity of the scenes in which waste dumps appear and present a diverse nature of the objects that compose a waste deposit. When observed from above waste dumps appear as complex arrangements of objects of different shapes, sizes, and orientations. The dataset consists of more than 11,700 images from three different sources and contains annotations at different granularities: Binary labels: images are classified based on the presence or absence of waste. Multi-class multi-label: a subset of images is annotated based on the presence of specific waste objects. Weakly-supervised localization: a subset of images is annotated with ground truth segmentation masks surrounding relevant waste objects. Visit the Github page for scripts and instructions on how to use this dataset: https://github.com/rnt-pmi/AerialWaste Visit the Dataset Page for more information: https://aerialwaste.org/ License Creative Commons CC BY licensing scheme (see LICENSE file). The usage of Google Imagery must respect the Google Earth terms and conditions [https://about.google/brand-resource-center/products-and-services/geo-guidelines/]. Acknowledgements This work was partially funded by the European Union’s Horizon Europe project PERIVALLON – Protecting the EuRopean terrItory from organised enVironmentAl crime through inteLLigent threat detectiON tools, under grant agreement no. 101073952.