| ID |
505 |
| Title |
Automatic segmentation of seismic vertical sections for subsurface interface identification |
| Creator |
Zan Wang |
| Subject |
Segmented seismic image data |
| Publisher |
Xiukuan Zhao |
| Description |
Delineating subsurface interfaces is a crucial step in site selection and characterization for various subsurface applications, such as the geologic carbon sequestration and hydrocarbon exploration and production. We developed an unsupervised learning method with deep fully convolutional networks (FCNs) for rapid subsurface interface identification on seismic vertical sections. We evaluate the performance of the FCN model using a real seismic dataset, which is publicly available at the CO2 DataShare website (https://co2datashare.org/dataset/smeaheia-dataset/resource/705d84fe-3054-4ab4-951b-c045782078fb). We extracted two inlines (inline numbers 1079 and 1406) and two crosslines (crossline numbers 2978 and 3750) in the 3D post-stack time migrated seismic dataset and used the four seismic amplitude images as inputs of the FCN model. The detailed description of our training method can be found in the main text of our manuscript. The code implementing the proposed unsupervised learning method for subsurface interface identification and the resulting segmentation results for the four input 2D seismic vertical sections are provided in this Data Repository. |
| Contributor |
Shengwen Qi, Youshan Liu, Bowen Zheng, Peng Sun and Yu Han |
| Date |
June 2022 - April 2023 |
| Type |
The seismic amplitude images are automatically segmented into three categories, denoted by different integers (i.e., the class labels) for the background, the peaks of seismic reflection events and the troughs of seismic reflection events, respectively. |
| Format |
.npy file format, available in Python |
| URL |
http://www.geophys.ac.cn/ArticleData/20230616seismic_interface.zip |
| DOI |
10.12197/2023GA006 |
| Source |
|
| Language |
eng |
| Relation |
|
| Coverage |
|
| Rights |
Institute of Geology and Geophysics, Chinese Academy of Sciences |
