Attain Dataset

Image 1 Image 2

The Attain (Artificial Intelligence in Transportation Lab) dataset establishes a new benchmark for pavement distress analysis through its comprehensive, high-quality standardized imagery. Its substantial size enables meaningful algorithm comparison and evaluation using real-world road surface data.

Attain_RSP_OS_v2.0

Images Taken With: Road Surface Profiler

Images View: Forward

Number of Images: 3350

Dimensions of Images: 1600× 1352

Color Model of Images: RGB

Annotation: Manually with polygonal segmentation

Labels Format: Text-based Format

Labeled Classes: Distress type

Distress:

  • Alligator Cracking(3796)

Attain_RSP_WS_v2.0

Images Taken With: Road Surface Profiler

Images View: Forward

Number of Images: 3350

Dimensions of Images: 1600× 1352

Color Model of Images: RGB

Annotation: Manually with polygonal segmentation

Labels Format: Text-based Format

Labeled Classes: Distress type and severity

Distress:

  • Alligator Cracking: Low Severity(901) - High Severity(2895)

Attain_RSP_OS_v1.0

Images Taken With: Road Surface Profiler

Images View: Forward

Number of Images: 13588

Dimensions of Images: 1600× 1352

Color Model of Images: RGB

Annotation: Manually with bonding box

Labels Format: Text-based Format

Labeled Classes: Background and distress type

Distress:

  • Bleeding(1885)
  • Bumps and Sags(857)
  • Corrugation(194)
  • Manhole(736)
  • Patching(3440)
  • Pothole(2074)
  • Rutting(14509)
  • Shoving(1556)

Attain_RSP_WS_v1.0

Images Taken With: Road Surface Profiler

Images View: Forward

Number of Images: 13041

Dimensions of Images: 1600× 1352

Color Model of Images: RGB

Annotation: Manually with bonding box

Labels Format: Text-based Format

Labeled Classes: Background and distress type and severity

Distress:

  • Bleeding: Low Severity(10) - Medium Severity(1817) - High Severity(58)
  • Bumps and Sags: Low Severity(588) - High Severity(269)
  • Corrugation: Low Severity(22) - High Severity(172)
  • Patching: Low Severity(2265) - High Severity(1175)
  • Pothole: Low Severity(1680) - High Severity(394)
  • Rutting: Low Severity(4592) - High Severity(9917)
  • Shoving: Low Severity(1298) - High Severity(258)

Attain_SMP_WS_v2.0

Images Taken With: Smartphone

Images View: Forward

Number of Images: 847

Dimensions of Images: 640× 640 - 1479x 508

Color Model of Images: RGB

Annotation: Manually with bonding box

Labels Format: XML-based Format

Labeled Classes: Background and distress type and severity

Distress:

  • Alligator Cracking: Low Severity(2489) - High Severity(215)
  • Block Crack: Low Severity(44) - High Severity(14)
  • Faded Marking: Low Severity(733) - High Severity(630)
  • Lane Shoulder Drop-off: High Severity(348)
  • Linear Crack: Low Severity(3494) - High Severity(702)
  • Patch and Utility Cut: Low Severity(705) - High Severity(41)
  • Pothole: Low Severity(257) - High Severity(58)
  • Raveling: Low Severity(329) - High Severity(11)
  • Weathering: Low Severity(1106) - High Severity(534)

Attain_SMP_OS_v1.0

Images Taken With: Smartphone

Images View: Forward

Number of Images: 637

Dimensions of Images: 640× 640

Color Model of Images: RGB

Annotation: Manually with bonding box

Labels Format: Text-based Format

Labeled Classes: Background and distress type

Distress:

  • Alligator crack(198)
  • Linear crack(720)

Attain_SMP_WS_v1.0

Images Taken With: Smartphone

Images View: Forward

Number of Images: 809

Dimensions of Images: 640× 640

Color Model of Images: RGB

Annotation: Manually with bonding box

Labels Format: Text-based Format

Labeled Classes: Background and distress type and severity

Distress:

  • Alligator Crack: Low Severity(808) - Medium Severity(479) - High Severity(62)
  • Block Crack: Low Severity(3)
  • Faded Marking: Low Severity(61) - High Severity(33)
  • Linear Crack: Low Severity(3731) - High Severity(372)
  • Manhole: Low Severity(484) - High Severity(20)
  • Patch: Low Severity(129)
  • Pothole: Low Severity(158) - High Severity(52)
  • Raveling: Low Severity(20)
  • Weathering: Low Severity(626) - High Severity(95)

Download Instructions
  1. Fill out the form below
  2. Your request will be processed by the Attain lab in 2 working days.
  3. An email will be sent to you including a link to download the requested file upon request acceptance.
RSP
Citation
Attain_RSP_OS_v2.0

Sholevar, N., Golroo, A., & Esfahani, S. R. (2022). Machine learning techniques for pavement condition evaluation. Automation in Construction, 136, 104190.

Attain_RSP_WS_v2.0

Sholevar, N., Golroo, A., & Esfahani, S. R. (2022). Machine learning techniques for pavement condition evaluation. Automation in Construction, 136, 104190.

Attain_RSP_OS_v1.0

Valipour, P. S., Golroo, A., Kheirati, A., Fahmani, M., & Amani, M. J. (2024). Automatic pavement distress severity detection using deep learning. Road Materials and Pavement Design, 25(8), 1830-1846.

Attain_RSP_WS_v1.0

Fahmani, M., Golroo, A., & Sedighian-Fard, M. (2024). Deep learning-based predictive models for pavement patching and manholes evaluation. International Journal of Pavement Engineering, 25(1), 2349901.

Attain_SMP_WS_v2.0

-

Attain_SMP_OS_v1.0

-

Attain_SMP_WS_v1.0

-

Request Dataset