Photogranulometry - Dataset of soil images with corresponding particle size distributions

  • Federated Research Data Repository / dépôt fédéré de données de recherche
  • Thomas Plante St-Cyr (Creator)
  • François Duhaime (Contributor)
  • Jean-Sébastien Dubé (Creator)
  • Simon Grenier (Creator)
  • Thien-Kim Leduc Tong (Contributor)
  • Lianne Fray (Contributor)
  • Thomas Plante St-Cyr (Contributor)
  • WSP Canada (Contributor)
  • François Duhaime (Contributor)
  • Jean-Sébastien Dubé (Contributor)
  • Simon Grenier (Contributor)

Dataset

Description

This file contains a high-resolution image dataset of 321 varied geotechnical soil samples, each paired with laboratory-measured particle size distribution (PSD) data from mechanical sieving and hydrometer test. The 12,714 images were taken in both moist and dry states in a standardized top-view conditions. The samples were collected in the Montreal, Canada region and they cover a wide range of granular textures. Each image has a resolution of 8192×5464 pixels (44.76 MP) with a scale of approximately 39.4 µm/pixel. This dataset was collected to support the development and validation of convolutional neural networks (CNNs) for optical grain size analysis in geotechnical engineering.

Cite this