Lithology dataset

The dataset includes core images used in the paper “Automated lithology classification from drill core images using convolutional neural networks” https://doi.org/10.1016/j.petrol.2020.107933. It includes 2 x 2 cm core images created from core tray images provided by the Geological Survey of South Australia, available on http://www.energymining.sa.gov.au/minerals/geoscience/geoscientific_data/hylogger. It includes images used for training, validating, and testing the CNN models for four categories: sandstone, limestone, shale, and a ‘garbage’ category that contains empty tray areas and non-core objects such as and wooden/metallic objects in the tray images.