This repository contains a set of workflows used to generate a Volve-alike synthetic dataset for testing of seismic processing and imaging algorithms.
NOTE: due to their large size, the dataset cannot be shared directly in this repository. You can however download it from this Zenodo link and place its content into the Data directory.
This repository is organized as follows:
- Inversion: directory containing a jupyter notebook to perform poststack inversion and save the velocity model used for modelling.
- Modelling: directory containing Madagascar scripts for modelling.
- Processing: directory containing Python notebooks for processing.
- Imaging: directory containing Madagascar scripts and Python codes for imaging.
- Visualization: directory containing Python used to visualize the generated data.
- Data: directory containing all the datasets that we have generated in .npz and/or binary format.
- Figures: directory containing all figures created in the Python scripts.
A visual description of the various components is provided in the figure at the top of the README fine.
As part of this project, we have used 2 main pieces of software:
- Python: used to create the sharp velocity model by seismic inversion, process seismic data (i.e., up/down sepatation and deconvolution), imaging and visualization. In all cases we rely on tools provided by the PyLops framework for inverse problems.
- Madagascar: used for modelling and imaging. More specifically, we use a vector-acoustic (i.e., first-order staggered grid acoustic wave equation) FD modelling code for modelling, and a acoustic FD modelling code for imaging.
We envision two levels of users: basic and advanced.
Basic users simply want to access the data that we have already modelled, alongside the velocity model and images. Such users should head over to
the Visualization directory where they will be able to see how the different files created in this repository can be loaded and visualized in Python.
Advanced users are perhaps interested to create their own data using more advanced physics (e.g., elastic, attenuation, anisotropy) or look into the details of how our velocity model and data have been created. In this case, we invite users to refer to the above figure to understand which directory contains the required script to create our derivative work.