A Diverse Super Resolution Dataset of Digital Rocks (DeepRock-SR): Sandstone, Carbonate, and Coal

A Diverse Super Resolution Dataset of Digital Rocks (DeepRockSR): Sandstone, Carbonate, and Coal.

This dataset contains an organised and processed collection of greyscale digital rock images for the purpose of image super resolution training. In total, there are 12,000 2D images and 3,000 3D volumes contained in this dataset.

Sandstone:

Bentheimer Sandstone 1 https://www.digitalrocksportal.org/projects/211
Bentheimer Sandstone 2 https://www.digitalrocksportal.org/projects/135
Berea Sandstone https://www.digitalrocksportal.org/projects/135
Leopard Sandstone https://www.digitalrocksportal.org/projects/135
Gildehauser Sandstone https://www.digitalrocksportal.org/projects/134
Wilcox Tight Sandstone https://www.digitalrocksportal.org/projects/6

Carbonate:

Estaillades Carbonate https://www.digitalrocksportal.org/projects/58
Savonnieres Carbonate https://www.digitalrocksportal.org/projects/72
Massangis Carbonate https://www.digitalrocksportal.org/projects/73

Coal:

Sheared Coal https://www.digitalrocksportal.org/projects/21
Naturally Fractured Coal https://www.digitalrocksportal.org/projects/20

The dataset is organised in a similar fashion to the DIV2K dataset : (https://data.vision.ee.ethz.ch/cvl/DIV2K/). To the knowledge of the authors, this dataset is the first of its kind to be used for the purpose of benchmarking Super Resolution algorithms in digital rock images. This dataset has been used thus far to test several Super Resolution Convolutional Neural Networks (SRCNN). Results are described in the Related Publications tab on this page.

Please cite both this dataset (use its DOI) as well as the related journal publication.

Data overview:
The dataset is divided into:
• 2D and 3D folders
• Sandstone, Carbonate, and Coal Datasets
• Combined shuffled Datasets
• High Resolution, 2x and 4x downsampled datasets using the Matlab imresize and imresize3 functions with default settings and randomised settings
• 2D HR images measure 500x500, cropped samples from the centre of the original cylindrical images, saved as PNG files
• 9600 2D images are available for training, 1200 for validation, and 1200 for testing.
• 3D HR volumes measure 100x100x100, cropped from the centre of the original cylindrical images, saved as octave and scipy readable MAT files (each array name is 'temp')
• 2400 3D volumes are available for training, 300 for validation, and 300 for testing.

The folder structure is as follows:
• DeepRockSR/ – the parent folder
• DeepRockSR/DeepRockSR-2D/ – folder for all 2D data
• DeepRockSR/DeepRockSR-2D/sandstone2D – folder for 2D sandstone Data
• DeepRockSR/DeepRockSR-2D/carbonate2D– folder for 2D carbonate Data
• DeepRockSR/DeepRockSR-2D/coal2D– folder for 2D coal Data
• DeepRockSR/DeepRockSR-2D/shuffled2D – folder for 2D shuffled Data

and similarly for the 3D dataset, under the exact same structure.

Within each carbonate, sandstone, and shuffled folder, the structure is as follows, using the 2D shuffled folder as example:
• /shuffled2D_train_HR/ – 0001.png, 0002.png, ..., 0800.png train HR images
• /shuffled2D_train_LR_default_X2/ -- 0001x2.png, 0002x2.png, ..., 0800x2.png train LR images, downscale factor x2
• /shuffled2D_train_LR_default_X4/ -- 0001x4.png, 0002x4.png, ..., 0800x4.png train LR images, downscale factor x4
• /shuffled2D_train_LR_unknown_X2/ -- 0001x2.png, 0002x2.png, ..., 0800x2.png train LR images, downscale factor x2
• /shuffled2D_train_LR_unknown_X4/ -- 0001x4.png, 0002x4.png, ..., 0800x4.png train LR images, downscale factor x4
• Validation
• /shuffled2D_valid_HR/ -- 0801.png, 0802.png, ..., 0900.png validation HR images
• /shuffled2D_valid_LR_default_X2/ -- 0801x2.png, 0802x2.png, ..., 0900x2.png train LR images, downscale factor x2
• /shuffled2D_valid_LR_default_X4/ -- 0801x4.png, 0802x4.png, ..., 0900x4.png train LR images, downscale factor x4
• /shuffled2D_valid_LR_unknown_X2/ -- 0801x2.png, 0802x2.png, ..., 0900x2.png train LR images, downscale factor x2
• /shuffled2D_valid_LR_unknown_X4/ -- 0801x4.png, 0802x4.png, ..., 0900x4.png train LR images, downscale factor x4
• Testing
• /shuffled2D_test_HR/ -- 0901.png, 0902.png, ..., 1000.png test HR images
• shuffled2D_test_LR_default_X2/ -- 0901x2.png, 0902x2.png, ..., 1000x2.png train LR images, downscale factor x2
• shuffled2D_test_LR_default_X4/ -- 0901x4.png, 0902x4.png, ..., 1000x4.png train LR images, downscale factor x4
• shuffled2D_test_LR_unknown_X2/ -- 0901x2.png, 0902x2.png, ..., 1000x2.png train LR images, downscale factor x2
• shuffled2D_test_LR_unknown_X4/ -- 0901x4.png, 0902x4.png, ..., 1000x4.png train LR images, downscale factor x4

多样化数字岩石超分辨率数据集（DeepRockSR）：砂岩、碳酸盐岩与煤岩

本数据集包含经整理与预处理的灰度数字岩石图像集合，用于图像超分辨率模型训练。数据集总计包含12000张二维图像与3000个三维体数据。

砂岩：
本特海姆砂岩1 https://www.digitalrocksportal.org/projects/211
本特海姆砂岩2 https://www.digitalrocksportal.org/projects/135
贝雷砂岩 https://www.digitalrocksportal.org/projects/135
豹斑砂岩 https://www.digitalrocksportal.org/projects/135
吉尔道泽砂岩 https://www.digitalrocksportal.org/projects/134
威尔科克斯致密砂岩 https://www.digitalrocksportal.org/projects/6

碳酸盐岩：
埃斯塔亚德碳酸盐岩 https://www.digitalrocksportal.org/projects/58
萨沃尼耶尔碳酸盐岩 https://www.digitalrocksportal.org/projects/72
马桑吉碳酸盐岩 https://www.digitalrocksportal.org/projects/73

煤岩：
剪切煤岩 https://www.digitalrocksportal.org/projects/21
天然裂缝煤岩 https://www.digitalrocksportal.org/projects/20

本数据集的组织形式与DIV2K数据集（https://data.vision.ee.ethz.ch/cvl/DIV2K/）类似。据作者所知，本数据集是首个用于数字岩石图像超分辨率算法基准测试的同类数据集。目前该数据集已被用于测试多款超分辨率卷积神经网络（Super Resolution Convolutional Neural Networks，SRCNN），相关研究结果可参见本页面的「相关出版物」标签页。

请同时引用本数据集（使用其DOI）与相关期刊论文。

数据概览：
本数据集划分如下：
• 二维与三维数据文件夹
• 砂岩、碳酸盐岩与煤岩数据集
• 混合打乱数据集
• 高分辨率（High Resolution，HR）数据集，以及采用Matlab的imresize与imresize3函数、以默认参数与随机参数下采样得到的2倍、4倍低分辨率数据集
• 二维高分辨率图像尺寸为500×500，为从原始圆柱形图像中心裁剪得到的样本，保存为PNG格式文件
• 二维图像共9600张用于训练、1200张用于验证、1200张用于测试
• 三维高分辨率体数据尺寸为100×100×100，为从原始圆柱形图像中心裁剪得到的样本，保存为可被Octave与Scipy读取的MAT文件（数组名均为'temp'）
• 三维体数据共2400个用于训练、300个用于验证、300个用于测试

文件夹结构如下：
• DeepRockSR/ — 根文件夹
• DeepRockSR/DeepRockSR-2D/ — 所有二维数据文件夹
• DeepRockSR/DeepRockSR-2D/sandstone2D — 二维砂岩数据文件夹
• DeepRockSR/DeepRockSR-2D/carbonate2D — 二维碳酸盐岩数据文件夹
• DeepRockSR/DeepRockSR-2D/coal2D — 二维煤岩数据文件夹
• DeepRockSR/DeepRockSR-2D/shuffled2D — 二维打乱混合数据文件夹
三维数据集的文件夹结构与此完全一致。

以二维打乱混合文件夹为例，各砂岩、碳酸盐岩、煤岩与打乱混合文件夹内的结构如下：
• /shuffled2D_train_HR/ — 0001.png、0002.png、……、0800.png，训练集高分辨率图像
• /shuffled2D_train_LR_default_X2/ — 0001x2.png、0002x2.png、……、0800x2.png，训练集低分辨率图像，下采样倍率为2倍（默认参数）
• /shuffled2D_train_LR_default_X4/ — 0001x4.png、0002x4.png、……、0800x4.png，训练集低分辨率图像，下采样倍率为4倍（默认参数）
• /shuffled2D_train_LR_unknown_X2/ — 0001x2.png、0002x2.png、……、0800x2.png，训练集低分辨率图像，下采样倍率为2倍（随机参数）
• /shuffled2D_train_LR_unknown_X4/ — 0001x4.png、0002x4.png、……、0800x4.png，训练集低分辨率图像，下采样倍率为4倍（随机参数）
• 验证集
• /shuffled2D_valid_HR/ — 0801.png、0802.png、……、0900.png，验证集高分辨率图像
• /shuffled2D_valid_LR_default_X2/ — 0801x2.png、0802x2.png、……、0900x2.png，验证集低分辨率图像，下采样倍率为2倍（默认参数）
• /shuffled2D_valid_LR_default_X4/ — 0801x4.png、0802x4.png、……、0900x4.png，验证集低分辨率图像，下采样倍率为4倍（默认参数）
• /shuffled2D_valid_LR_unknown_X2/ — 0801x2.png、0802x2.png、……、0900x2.png，验证集低分辨率图像，下采样倍率为2倍（随机参数）
• /shuffled2D_valid_LR_unknown_X4/ — 0801x4.png、0802x4.png、……、0900x4.png，验证集低分辨率图像，下采样倍率为4倍（随机参数）
• 测试集
• /shuffled2D_test_HR/ — 0901.png、0902.png、……、1000.png，测试集高分辨率图像
• shuffled2D_test_LR_default_X2/ — 0901x2.png、0902x2.png、……、1000x2.png，测试集低分辨率图像，下采样倍率为2倍（默认参数）
• shuffled2D_test_LR_default_X4/ — 0901x4.png、0902x4.png、……、1000x4.png，测试集低分辨率图像，下采样倍率为4倍（默认参数）
• shuffled2D_test_LR_unknown_X2/ — 0901x2.png、0902x2.png、……、1000x2.png，测试集低分辨率图像，下采样倍率为2倍（随机参数）
• shuffled2D_test_LR_unknown_X4/ — 0901x4.png、0902x4.png、……、1000x4.png，测试集低分辨率图像，下采样倍率为4倍（随机参数）