Applied Diffraction Imaging: Conventional vs. Machine Learning Approach
|Title:||Applied Diffraction Imaging: Conventional vs. Machine Learning Approach||Authors:||Lowney, Brydon||Permanent link:||http://hdl.handle.net/10197/12859||Date:||2021||Online since:||2022-05-05T16:02:59Z||Abstract:||Diffractions are oft overlooked in favour of specular reflections for seismic imaging. Diffractions, however, are formed by objects and discontinuities which are comparable or smaller than the wavelength. Therefore, if the diffractions can be imaged, these objects and discontinuities can be directly imaged. Said features are geologically noteworthy as structural and stratigraphic features of interest which may directly affect hydrocarbon migration, flow, and trapping. While straightforward in theory, in application separating the diffractions from the wavefield is a complex and intricate task. This thesis aims to discuss what makes separation so problematic and attempts to address various methods for allaying some of these issues. Additionally, this thesis intends to tackle the real benefits diffraction imaging can add to a conventional seismic image. To achieve these goals, existing separation techniques are analysed in pre-migration and post-migration domains. These analytical methods leave a volume which contains both diffractions and noise and require additional inputs such as a dip field. Ergo, any errors present in the dip field are incorporated into the diffraction image, diminishing its quality. In this thesis, adaptations to existing methods have been proposed as well as a novel method in the pre-migration domain. These new methods incorporate common geophysical techniques to obtain a cleaner separation. The novel method, histogram-based separation, is an analytical curve-fitting technique to identify diffractions without the presence of noise. Deep learning has also been integrated with the methods. Firstly, with pre-migration separation deep learning has been used to automatically identify and separate diffractions, obtaining a cleaner separation in a faster time. Another neural network is also trained which can apply a post-migration-based diffraction imaging scheme directly on stacked, migrated, seismic data, something hitherto impossible. These methods have been applied to seismic and GPR data highlighting the benefits of diffraction imaging.||Type of material:||Doctoral Thesis||Publisher:||University College Dublin. School of Earth Sciences||Qualification Name:||Ph.D.||Copyright (published version):||2021 the Author||Keywords:||Diffractions; Seismic; Deep learning||Language:||en||Status of Item:||Peer reviewed||This item is made available under a Creative Commons License:||https://creativecommons.org/licenses/by-nc-nd/3.0/ie/|
|Appears in Collections:||Earth Sciences Theses|
Show full item record
If you are a publisher or author and have copyright concerns for any item, please email email@example.com and the item will be withdrawn immediately. The author or person responsible for depositing the article will be contacted within one business day.