Selected Opportunities for the Semiautomatic Analysis of Spectrally Decomposed Seismic Data
Master Degree Thesis
by Andreas Johnsen Lind (ali054uib.no)
supervised by Helwig Hauser
Seismic imaging/visualization of the subterranean is an important part in the process of identifying and extracting hydrocarbons such as oil and gas. One promising method that is used to assist in visualizing seismic reflection data is spectral decomposition. Spectral decomposition is a contrast enhancing technique that helps in identifying subterranean features.
This thesis presents a set of new opportunities for visualizing spectrally decomposed seismic based on techniques known from visualization and image Segmentation. In particular it focuses on presenting a new technique based on Interactive Visual Analysis (IVA) and feature extraction based image segmentation. In addition to the visualization and segmentation techniques, pre-processing steps for making the data more suitable for visualization are suggested. Some of the pre-processing techniques are well known from the image processing field.