Road Traversing Knowledge

The Road Transverse Knowledge (RTK) dataset was specially designed to meet the emerging country infrastructure. It sums 701 densely annotated images of low-resolution quality (352x288), split into 561 and 140 samples for the training and the validation sets, respectively. The dataset presents 12 classes of road surfaces like asphalt, paved and unpaved; damages like potholes, water puddles, and cracks; and road signs like speed bumps, cat's eyes, and markings. The images were captured in good weather and daylight conditions in the country areas of Santa Catarina in Brazil.