IPython notebook for the Chevron SEG 2014 FWI Benchmark Dataset

This notebook is designed to make it straightforward to get started with the 2014 full-waveform inversion (FWI) blind-test dataset provided by the SEG and Chevron. Please feel free to remix and redistribute this notebook, under the terms of the license.

View on IPython Notebook Viewer

Requirements

• IPython with notebook support
• matplotlib
• numpy
• scipy
• pygeo segyread.SEGYFile (included)

Setup

Most Linux distributions should make the required packages available automatically. On Ubuntu, for example, run

apt-get install ipython-notebook python-matplotlib python-scipy

To run the notebook, unpack the dataset to the data subdirectory (or you may symlink the data and/or modify the first code block of the notebook). Change to the project directory and run

ipython notebook

To simplify installation on a Mac OS or Windows machine you may want to consider using Enthought's Canopy distribution or Continuum Analytics's Anaconda distribution, which contain all of the required packages.

Notes

This notebook should work on Wakari with only minor modifications; you will need to remove the lines

%pylab
%matplotlib inline

These can be found in the third code box of the notebook, under General Setup. The full dataset fits within Wakari's free plan, and appears to run without difficulty.

License information

This IPython notebook is licensed under Creative Commons Attribution-ShareAlike 2.5 Canada. The SEG 2014 Chevron Full Waveform Inversion Synthetic dataset is licensed separately; see the license agreement included with the data. The SEG-Y library used is segyread.SEGYFile from pygeo, and is distributed under the GNU Lesser GPL.