Title:  Computational Seismic Interpretation using Attention Models, Texture Dissimilarity, and Learning

Committee:

Dr. Ghassan AlRegib, ECE, Chair , Advisor

Dr. James McClellan, ECE

Dr. David Anderson, ECE

Dr. Gordon Stuber, ECE

Dr. Zhigang Peng, EAS

Abstract:

The objective of this research is to automate the process of structural interpretation of migrated seismic data. Establishing a close correspondence between seismic interpretation and area geology is a key prerequisite of any successful exploration and production project. To extract useful information from seismic volumes, interpreters manually delineate important geological structures such as salt domes, faults, and anticlines that contain hints about petroleum and gas reservoirs. These structures, which typically span over several square kilometers are delineated based on the correlation, changes in illumination, intensity, contrast, and the texture of seismic data. Manual delineation of such features is extremely time consuming and labor intensive and unfortunately, very few tools are available for automatic seismic interpretation. In this dissertation research, we propose novel seismic attributes based on visual-attention theory, the modeling of human visual system, and machine learning to quantify changes and highlight geological features in a three-dimensional space. To automate the process of seismic interpretation, we develop interpreter-assisted, fully-, and semi-automated workflows that are interactive and easy-to-use for the delineation of important geological structures within seismic volumes. Experimental results on real and synthetic seismic datasets show that our proposed algorithms outperform state-of-the-art methods for seismic interpretation. The proposed research is computationally inexpensive and is expected to not only reduce the time for seismic interpretation but also become a handy tool in the interpreter’s toolbox for detecting and delineating important geological structures.