search

thilowrona / seismic_deep_learning

A couple of python scripts to interpret geological structures from geophysical images using deep learning
GNU Affero General Public License v3.0
198 stars 74 forks source link

Just a little request for an additional notebook #1

Closed ivan-marroquin closed 3 years ago

ivan-marroquin commented 3 years ago

Hi there!

Many thanks for making available these tutorials. I was hoping that you could add another notebook showing how to import seismic data, scale amplitude range, and same inlines (or crosslines) to a tiff file.

Kind regards,

Ivan

thilowrona commented 3 years ago

Hi Ivan,

thanks a lot for your message! Loading seismic data (usually stored in .segy format) can be a bit tricky, because information, such as number of traces, inlines and crosslines, are not always stored in the same way, so you probably need to explore your seismic files a bit.

The best tools I've found for loading seismic data in Python are segyio and segpy. They have plenty of examples for 2-D and 3-D data. There is also an example for 3-D data in Tutorial-3 In [6]. This also shows how to scale your data.

Finally, I can recommend Pythontool-pro by Cegal, which allows you to transfer all kinds of data (seismics, interpretations) between Python and Petrel - a great tool for machine learning on seimsic data.

To save seismic data as tiff, I have converted it from an array (NumPy) to an image (Pillow) and then saved it.

Hope any of these are useful!

All the best, Thilo

ivan-marroquin commented 3 years ago

Hi @thilowrona

Many thanks for your prompt answer and for sharing all this information.

Ivan