4Seasons is a dataset about seasonal and challenging perception conditions in autonomous driving. The data was collected in different scenes, different weather conditions, and different lighting conditions, such as day and night. The dataset contains images of more than 350 kilometers of distance in 9 environments, ranging from multi-storey car parks in cities (including tunnels) to rural areas and highways.

The dataset can be used to study visual mapping, global location recognition, and map-based relocalization tracking.