High Resolution Post-Stack Geophysical Processing using Principal Component Analysis
Scheevel Geo Technologies has developed a suite of Post-stack resolution enhancement processing tools. These tools function using statistical and neural network techniques. Our methodologies are primarily based on simple, but powerful statistical analysis methods. The bulk of our methods leverage the time-tested statistical approach, Primcipal Component Analysis (PCA). Our implementation of PCA is customized for application to seismic data and makes use of windowed samples of amplitude (or other seismic attributes) in order to successfully deconvolve the underlying stratigraphic architecture from raw seismic signal.
We have published on use of this technique to do reservoir predictions. Click on this link to downlaod a pdf on this subject.
Addtional examples of use of our techniques are shown in the follwoing links.
1. Click here to see the application of the technique to create high resolution facies clasification of 3D amplitude data.
2. Click here to see experimetal results using both PCA and a general regression neural network to make pseudo high resolution property predictions from raw amplitude data.
3. We also have developed high resolution seismic-facies pattern prediction filters that can further deconvolve stratigraphic features. An example is shown below. Original Amplitude is on the left. The image on the right is the high resolution pattern match. Well data is overposted to illustrate value of addtional resolution for purposes of correlation. Click here to download a detailed public presentation of this technology shown March 27, 2009 at the RMAG/DGS 15th annual 3D Seismic Symposium in Denver Colorado. The file is a winzip file that contains a graphical presentation and a short captioned movie (wmv) which summarizes the processing. File size is 12Mb.