Ahmad Mustafa, Motaz Alfarraj, and Ghasssan AlRegib

This repository includes the codes for the paper:

'Estimation of Acoustic Impedance from Seismic Data using Temporal Convolutional Network', Expanded Abstracts of the SEG Annual Meeting , San Antonio, TX, Sep. 15-20, 2019. [PDF]

The code has been built in Python 3.7.0 and requires the following dependencies before it can be run:

numpy 1.15.1
matplotlib 3.0.0
pytorch 0.4.1
seaborn 0.9.0
segyio 1.7.0
tensorboard 1.13.1
tensorboardx 1.7
tensorflow 1.13.1
pandas 0.24.2
scipy 1.1.0

In exploration seismology, seismic inversion refers to the process of inferring physical properties of the subsurface from seismic data. Knowledge of physical properties can prove helpful in identifying key structures in the subsurface for hydrocarbon exploration. In this work, we propose a workflow for predicting acoustic impedance (AI) from seismic data using a network architecture based on Temporal Convolutional Network by posing the problem as that of sequence modeling. The proposed workflow overcomes some of the problems that other network architectures usually face, like gradient vanishing in Recurrent Neural Networks, or overfitting in Convolutional Neural Networks. The proposed workflow was used to predict AI on Marmousi 2 dataset with an average r² coefficient of 91% on a hold-out validation set.

The repository contains all the data needed to run the codes. Clone the repo to an appropriate directory on your machine. Afterwards, use a dedicated python IDE like Spyder or PyCharm to view and execute the train.py file. Alternatively, you may run the codes from the command line as follows:

cd <project root directory>
python train.py --no_wells 12 --epochs 900 --data_flag <marmousi or seam>

If you have found our code and data useful, we humbly request you to cite our work.

@inbook{doi:10.1190/segam2019-3216840.1,
author = {Ahmad Mustafa and Motaz Alfarraj and Ghassan AlRegib},
title = {Estimation of acoustic impedance from seismic data using temporal convolutional network},
booktitle = {SEG Technical Program Expanded Abstracts 2019},
chapter = {},
pages = {2554-2558},
year = {2019},
doi = {10.1190/segam2019-3216840.1},
URL = {https://library.seg.org/doi/abs/10.1190/segam2019-3216840.1}
}

The arXiv preprint is available at: https://arxiv.org/abs/1906.02684

The code and the data are provided as is with no guarantees. If you have any questions, regarding the dataset or the code, you can contact me at ([email protected]), or even better, open an issue in this repo and we will do our best to help.