SeisGAN: Improving Seismic Image Resolution and Reducing Noise Using a Generative Adversarial Network

An application of generative adversarial networks to seismic data processing (resolution ehancement and denoising). This is a repository for the paper "SeisGAN: Improving Seismic Image Resolution and Reducing Noise Using a Generative Adversarial Network".

Example

Project Organization

Code

All training and test code are in the directory code.

Dataset

The three filed seismic images and the reconstructed images by our method are in application folder.The synthetic seismic data used for training can be obtained by visting the "https://www.kaggle.com/datasets/leilin1995/seisgan".

Dependencies

python 3.6.13
pytorch 1.9.1
torchvision 0.10.1
tqdm 4.62.3
scipy 1.5.4
numpy 1.19.5
h5py 3.1.0
pandas 1.1.5
PIL 8.4.0
matplotlib 3.3.4

Usage instructions

Download this project and build the dependency. Then use application_filed.py --test_data_path="your data path" --save_path="your save path of resed by GAN"

Citation

If you find this work useful in your research, please consider citing:

Lin L, Zhong Z, Cai C, et al. SeisGAN: Improving Seismic Image Resolution and Reducing Random Noise Using a Generative Adversarial Network[J]. Mathematical Geosciences, 2023: 1-27.

BibTex

@article{lin2023seisgan,
  title={SeisGAN: Improving Seismic Image Resolution and Reducing Random Noise Using a Generative Adversarial Network},
  author={Lin, Lei and Zhong, Zhi and Cai, Chuyang and Li, Chenglong and Zhang, Heng},
  journal={Mathematical Geosciences},
  pages={1--27},
  year={2023},
  publisher={Springer}
}

leilin1995 / SeisGAN-Improving-Seismic-Image-Resolution-and-Reducing-Noise

readme