| ID |
521 |
| Title |
ShaleSeg: The first deep-learning dataset and models for practical fracture segmentation of large-scale-shale CT images |
| Creator |
Yanfang Wu |
| Subject |
Large-scale shale CT scanning images |
| Publisher |
Xiukuan Zhao |
| Description |
The data includes two files in *.npy format. ‘data_all_pred_res.npy’ is the images of rock CT images. ‘data_label.npy’ is the label images. Two typical shapes of post-fracture shale, cubic shape (30 cm × 30 cm × 30 cm) and cylindrical shape (Ф50mm×100 mm), was scanned. A software VStudio software by Volume Graphics company was employed to manually label the fracture positions in the images and generate corresponding labels. The dataset was structured by 4313 images, which was structured as follows: 712 images (16.5%) featured severe circular artifacts and low contrast, while 3601 images (83.5%) exhibited notable variations in fracture width. |
| Contributor |
Juan Li, Shouding Li, Luqing Zhang, Jian Zhou, Jianming He, Xiao Li, Zhaobin Zhang |
| Date |
2022 |
| Type |
The resolution of cylindrical shape shale image was 0.129mm. The resolution of cubic shape shale image was 0.159mm. |
| Format |
*.npy, available in Python |
| URL |
http://www.geophys.ac.cn/ArticleData/20231023ShaleSeg.zip |
| DOI |
10.12197/2023GA022 |
| Source |
|
| Language |
eng |
| Relation |
|
| Coverage |
|
| Rights |
Institute of Geology and Geophysics, Chinese Academy of Sciences |
