apolonius22 / road-surface-detection-and-differentiation-considering-surface-damages

Semantic Segmentation in road surfaces, considering different pattern like asphalt, unpaved, potholes, etc.

Road surface detection and differentiation considering surface damages

The semantic segmentation GT for road surfaces contains 701 frames from RTK dataset. Classes are defined as follows:

• Background, everything being unrelated to the road surface;
• Asphalt, roads with asphalt surface;
• Paved, different pavements (eg.: Cobblestone);
• Unpaved, for unpaved roads;
• Markings, to the road markings;
• Speed-Bump, for the speed-bumps on the road;
• Cats-Eye, for the cats-eye found on the road, both on the side and in the center of the path;
• Storm-Drain, usually at the side edges of the road;
• Patch, for the various patches found on asphalt road;
• Water-Puddle, we use this class also for muddy regions;
• Pothole, for different types and sizes of potholes, no matter if they are on asphalt, paved or unpaved roads;
• Cracks, used in different road damages, like ruptures.

Citation:

@misc{rateke:2020.3,
   title = {Road surface detection and differentiation considering surface damages},
   author = {Thiago Rateke and Aldo von Wangenheim},
   journal={Autonomous Robots},
   year={2021},
   month={Jan},
   day={11},
   issn={1573-7527},
   doi={10.1007/s10514-020-09964-3},
   url={https://doi.org/10.1007/s10514-020-09964-3}
}

• Original Paper
• Step-by-step