Deep Learning for Seismic Resolution Enhancement, Enhancing Low-Resolution Legacy Data Using High Resolution Guidance

<a href="https://embed.stanford.edu/iframe/?url=https%3A%2F%2Fpurl.stanford.edu%2Frk466fq3383" class="su-underline">Show Content</a>

Abstract/Contents

Abstract: This study investigates the utilization of the CycleGAN neural network architecture to potentially enhance resolution in subsurface seismic imaging. While convolutional neural networks (CNNs) have been extensively applied for image enhancement, prior research mostly focused on increasing pixel resolution for detailed image enhancement. However, this thesis uniquely targets vertical and horizontal resolution issues in seismic data, often leading to undetected geological features largely due to insufficient frequency content in recorded seismic data. Enhancing subsurface seismic images involves significant resource costs, but this thesis showcases the potential to mitigate these costs by employing machine learning to enhance older-generation seismic data. Utilizing 3D seismic data from Saudi Arabia, including both older and higher resolution datasets, the research aims to address the frequency content and processing algorithms differences in historical datasets by training a CycleGAN model with a U-net as the CNN. The model aims to convert legacy seismic images to closely resemble higher-resolution datasets. Initial findings reveal noticeable resolution improvements, yet further refinement opportunities exist.

Type of resource	text
Publication date	November 28, 2023

Subject	Seismic
Subject	Deep learning (Machine learning)
Subject	Image processing
Genre	Text
Genre	Thesis

DOI	https://doi.org/10.25740/rk466fq3383
Location	https://purl.stanford.edu/rk466fq3383

Use and reproduction: User agrees that, where applicable, content will not be used to identify or to otherwise infringe the privacy or confidentiality rights of individuals. Content distributed via the Stanford Digital Repository may be subject to additional license and use restrictions applied by the depositor.
License: This work is licensed under a Creative Commons Attribution Non Commercial 4.0 International license (CC BY-NC).

Preferred citation: Al Qatari, H. (2024). Deep Learning for Seismic Resolution Enhancement, Enhancing Low-Resolution Legacy Data Using High Resolution Guidance. Stanford Digital Repository. Available at https://purl.stanford.edu/rk466fq3383. https://doi.org/10.25740/rk466fq3383.

Master's Theses, Doerr School of Sustainability

Loading usage metrics...