Pothole Detection Datasets

These datasets were created by Rui (Ranger) Fan on 16/08/2019

1. Publication

These datasets were published along with our IEEE Transactions on Image Processing (T-IP) paper "Pothole Detection Based on Disparity Transformation and Road Surface Modeling":

2. Dataset details

Each dataset includes:

• disp: disparity maps
• disp_tc: transformed disparity maps in color
• disp_tg: transformed disparity maps in gray-scale
• label: pixel-level pothole labels
• ptcloud: 3D point clouds
• rgb: RGB images

3. Demo code

Run demo.m to visualize our datasets.

4. Citation

Please cite the following papers when using these datasets:

@article{fan2019pothole,
  title={Pothole Detection Based on Disparity Transformation and Road Surface Modeling},
  author={Fan, Rui and Ozgunalp, Umar and Hosking, Brett and Liu, Ming and Pitas, Ioannis},
  journal={IEEE Transactions on Image Processing},
  year={2019},
  publisher={IEEE}
}

@article{fan2018road,
  title={Road surface 3D reconstruction based on dense subpixel disparity map estimation},
  author={Fan, Rui and Ai, Xiao and Dahnoun, Naim},
  journal={IEEE Transactions on Image Processing},
  volume={27},
  number={6},
  pages={3025--3035},
  year={2018},
  publisher={IEEE}
}

5. Note:

The following pothole detection results are not utilized for algorithm evaluation:

Dataset 2, frame ID: 6, 7, 8, 13, 14, 15, 21, 22, 23, 37, 38, 39.

4. Contact

e-mail: [email protected]