This dataset contains 11k annotated images of Polish roads, specially curated for the object detection task. The data was collected using car-mounted cameras on Polish roads, mainly in Krakow. The images capture a variety of scenes, including different road types and various lighting conditions (day and night).

The dataset contains the following 11 object categories: cars (vehicles without trailers), different traffic signs (traffic signs other than warning and prohibition signs, mainly information and command signs), green lights (green traffic lights for cars only; green lights for pedestrians are not annotated), motorcycles, pedestrians (people and cyclists), crosswalks (pedestrians crossing the street), prohibition signs (all prohibition signs), red lights (red lights for cars only; red lights for pedestrians are not annotated), speed limit signs (speed limit signs), trucks (vehicles with trailers), warning signs (warning signs).

Annotation process

Annotation was performed using Roboflow. A total of 2,000 images were manually annotated and another 9,000 images were generated through data augmentation. Annotation techniques applied included cropping, saturation, brightness, and exposure adjustments.