maksimbaz / cnn_vti_inversion

Determination of the elastic parameters of a VTI medium from sonic logging data using deep learning

cnn_vti_inversion

Determination of the elastic parameters of a VTI medium from sonic logging data using deep learning

---

Maksim Bazulin ([email protected]), Denis Sabitov, Marwan Charara

Codes prepared for Computers and Geosciences journal

---

Description

dataset10

The folder contains 10 shot gathers. This data can not be used for a proper neural network training but can be used for the testing.

dataset10/dataset_log_file_TD.txt

The text file that contains parameters of the medium related to the shot gathers, where columns denote respectively: file_number rho vp vs eps gamma delta

main_train.py

The script is devoted to the neural network construction and training.

main_test.py

The script is devoted to the neural network testing.

form_lists.py

The script is used for the list construction for the inputs and outputs for the neural network.

make_plot.ipynb

The script can be used to visualize shot gathers from dataset10 folder.

train_output

This is the folder, where you can find outputs of the main_train.py and main_test.py. This folder usually contains 'model.h5' - the file of the trained neural network weigths. This file is too large for uploading it, so do not hesitate to contact me if you need a sample of such a file.