Lab 17: Deep Learning with U-Net to Perform Seismic Inversion

CREWES Data Science Initiative will host on June 10th, at 5:30pm (MT) the learning lab 17. It will be the third lab of a series of four, focused on the solutions for seismic inversion using machine learning.
Geophysics in the Cloud was a competition with the goal to perform seismic inversion of rock attributes from seismic data with the use of well logs. It used open data (3D Poseidon from Australia) and the competitors needed to perform inversions for P-Impedance, S-Impedance, and Density. Well logs with DTC, DTS and RHOB are used for training and evaluation (two blin wells). After the data analysis, pre-processing, and feature engineering (presented during labs 15 and 16), now comes the modelling part. There are a large number of solutions to choose from. One of them is to use a U-Net deep learning model. This solution treats the seismic traces as 1-D images and uses a window of samples to perform local predictions. It has the advantage to consider each sample as dependent on its neighbours.
Zhan Niu is an MSc student at CREWES under the supervision of Dr. Daniel Trad and is a specialist in machine learning. He created a U-Net style deep learning model that inputs all the features created during the data engineering step, and outputs all the three targets at once, using the package Tensorflow.
Zhan will do a hands-on demonstration during the learning lab and will use the same dataset provided in the competition.
Zhan Niu's LinkedIn: / zhan-niu-8b53a9154
To learn more, visit our website:
www.crewes.org...
Follow us on Twitter: / crewes_uofc
Follow us on LinkedIn: / 4122656
Stay tuned for the next events at our MeetUp group: www.meetup.com...

[@sunjaysunjay6062]

Excellent great task greetings

[@BinhKieu82]

Hey, very nice Unet for seismic inversion slices! Congrats for the good matches for Vp, Vs. Correct me if im wrong, you only used extracted seismic attributes (16 attributes) a long well trajectory as inputs for the Unet? Have you used at least a seismic 2D line feeding in you model?
Can't wait to see your webinars about Geoscience this year, cheers!
BTW, if posible, could you share the github for the code please, thank you