{"id":{"repo_id":"alabama","oai_identifier":"oai:ir.ua.edu:123456789/14409"},"canonical_url":"https://search.dev.ndltd.org/etd/alabama/oai:ir.ua.edu:123456789/14409","repository":{"repo_id":"alabama","name":"University of Alabama","base_url":"https://ir-api.ua.edu/oai/request"},"display":{"title":"Seismic Data Processing and Interpretation via Deep Learning","abstract":"Deep learning (DL) algorithms are growing in popularity in seismic data processing and interpretation due to their efficiency and the ability to deal with non-linear problems. My dissertation focuses on developing new algorithms and workflows for seismic data processing and interpretation by using DL algorithms. The research of my dissertation concentrates on one topic in seismic data processing (Chapter 2) and two topics (Chapters 3 and 4) in seismic interpretation.Picking the first arrival times of shot gathers is a time-consuming but important procedure in seismic data processing. The first part of my dissertation (Chapter 2) uses deep learning algorithms to pick the first arrival times of shot gathers. First, the first arrival times of shot gathers are predicted using a U-Net deep learning architecture. Then, a workflow is developed to detect the seismic traces that were inaccurately predicted first arrival time. Next, an algorithm is proposed to pick the first arrival time for those seismic traces that have inaccurate picking of the first arrival time predicted by the U-Net deep learning architecture. Seismic horizon interpretation is one of the two main tasks in seismic structural interpretation. An experienced interpreter usually needs weeks or even months to interpret the horizons of a 3D seismic survey. The second part of my dissertation concentrates on accelerating the process of seismic horizon interpretation. Instead of adopting algorithms to automatically extract horizons, I propose to manually interpret horizons on a few key seismic sections and extract horizons under the constraints of manual interpretation.The third part of my dissertation focuses on seismic impedance inversion using physics-informed neural networks. The inputs of the neural networks are seismic amplitude, wavelet, and a low frequency model built using well logs. The output of the neural network is the seismic impedance that corresponds to the seismic trace. The optimization process is minimizing the difference between seismic trace and synthetic seismic waveforms computed from the output impedance log and input wavelet.","abstract_html":"Deep learning (DL) algorithms are growing in popularity in seismic data processing and interpretation due to their efficiency and the ability to deal with non-linear problems. My dissertation focuses on developing new algorithms and workflows for seismic data processing and interpretation by using DL algorithms. The research of my dissertation concentrates on one topic in seismic data processing (Chapter 2) and two topics (Chapters 3 and 4) in seismic interpretation.Picking the first arrival times of shot gathers is a time-consuming but important procedure in seismic data processing. The first part of my dissertation (Chapter 2) uses deep learning algorithms to pick the first arrival times of shot gathers. First, the first arrival times of shot gathers are predicted using a U-Net deep learning architecture. Then, a workflow is developed to detect the seismic traces that were inaccurately predicted first arrival time. Next, an algorithm is proposed to pick the first arrival time for those seismic traces that have inaccurate picking of the first arrival time predicted by the U-Net deep learning architecture. Seismic horizon interpretation is one of the two main tasks in seismic structural interpretation. An experienced interpreter usually needs weeks or even months to interpret the horizons of a 3D seismic survey. The second part of my dissertation concentrates on accelerating the process of seismic horizon interpretation. Instead of adopting algorithms to automatically extract horizons, I propose to manually interpret horizons on a few key seismic sections and extract horizons under the constraints of manual interpretation.The third part of my dissertation focuses on seismic impedance inversion using physics-informed neural networks. The inputs of the neural networks are seismic amplitude, wavelet, and a low frequency model built using well logs. The output of the neural network is the seismic impedance that corresponds to the seismic trace. The optimization process is minimizing the difference between seismic trace and synthetic seismic waveforms computed from the output impedance log and input wavelet.","abstract_has_math":false,"creators":["Pu, Yitao"],"institution":"University of Alabama Libraries","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Çemen, Ibrahim","Plattner, Alain","Zhu, Wei","Zeng, Hongliu"],"advisors":["Zhang, Bo"],"committee_chairs":[],"committee_members":[],"year":2024,"date_issued":"2024","date_published":"2024","updated_at":"2026-07-27T18:44:20Z","subjects":[],"languages":["en_US","English"],"rights":["All rights reserved by the author unless otherwise indicated."],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier.other","label":"Dc Identifier Other","values":["1085958"],"render_values":[{"text":"1085958","href":null,"code":true}]}]},"links":{"outbound_url":"https://ir.ua.edu/handle/123456789/14409","outbound_label":"Repository record","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Çemen, Ibrahim","Plattner, Alain","Zhu, Wei","Zeng, Hongliu"]},{"key":"dc:contributor.advisor","label":"Advisor","values":["Zhang, Bo"]},{"key":"dc:creator","label":"Author","values":["Pu, Yitao"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2024-09-17T16:18:44Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2024-09-17T16:18:44Z"]},{"key":"dc:date.issued","label":"Date","values":["2024"]},{"key":"dc:publisher","label":"Institution","values":["University of Alabama Libraries"]},{"key":"dc:type","label":"Dc Type","values":["thesis","text"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English"]},{"key":"dc:language.iso","label":"Language (ISO)","values":["en_US"]},{"key":"dc:rights","label":"Dc Rights","values":["All rights reserved by the author unless otherwise indicated."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.other","label":"Dc Identifier Other","values":["1085958"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://ir.ua.edu/handle/123456789/14409"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Electronic Thesis or Dissertation"]},{"key":"dc:description.abstract","label":"Abstract","values":["Deep learning (DL) algorithms are growing in popularity in seismic data processing and interpretation due to their efficiency and the ability to deal with non-linear problems. My dissertation focuses on developing new algorithms and workflows for seismic data processing and interpretation by using DL algorithms. The research of my dissertation concentrates on one topic in seismic data processing (Chapter 2) and two topics (Chapters 3 and 4) in seismic interpretation.Picking the first arrival times of shot gathers is a time-consuming but important procedure in seismic data processing. The first part of my dissertation (Chapter 2) uses deep learning algorithms to pick the first arrival times of shot gathers. First, the first arrival times of shot gathers are predicted using a U-Net deep learning architecture. Then, a workflow is developed to detect the seismic traces that were inaccurately predicted first arrival time. Next, an algorithm is proposed to pick the first arrival time for those seismic traces that have inaccurate picking of the first arrival time predicted by the U-Net deep learning architecture. Seismic horizon interpretation is one of the two main tasks in seismic structural interpretation. An experienced interpreter usually needs weeks or even months to interpret the horizons of a 3D seismic survey. The second part of my dissertation concentrates on accelerating the process of seismic horizon interpretation. Instead of adopting algorithms to automatically extract horizons, I propose to manually interpret horizons on a few key seismic sections and extract horizons under the constraints of manual interpretation.The third part of my dissertation focuses on seismic impedance inversion using physics-informed neural networks. The inputs of the neural networks are seismic amplitude, wavelet, and a low frequency model built using well logs. The output of the neural network is the seismic impedance that corresponds to the seismic trace. The optimization process is minimizing the difference between seismic trace and synthetic seismic waveforms computed from the output impedance log and input wavelet."]},{"key":"dc:format.medium","label":"Dc Format Medium","values":["electronic"]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Seismic Data Processing and Interpretation via Deep Learning"]}]}],"canonical_facts":{"dc:contributor":["Çemen, Ibrahim","Plattner, Alain","Zhu, Wei","Zeng, Hongliu"],"dc:contributor.advisor":["Zhang, Bo"],"dc:creator":["Pu, Yitao"],"dc:date.accessioned":["2024-09-17T16:18:44Z"],"dc:date.available":["2024-09-17T16:18:44Z"],"dc:date.issued":["2024"],"dc:description":["Electronic Thesis or Dissertation"],"dc:description.abstract":["Deep learning (DL) algorithms are growing in popularity in seismic data processing and interpretation due to their efficiency and the ability to deal with non-linear problems. My dissertation focuses on developing new algorithms and workflows for seismic data processing and interpretation by using DL algorithms. The research of my dissertation concentrates on one topic in seismic data processing (Chapter 2) and two topics (Chapters 3 and 4) in seismic interpretation.Picking the first arrival times of shot gathers is a time-consuming but important procedure in seismic data processing. The first part of my dissertation (Chapter 2) uses deep learning algorithms to pick the first arrival times of shot gathers. First, the first arrival times of shot gathers are predicted using a U-Net deep learning architecture. Then, a workflow is developed to detect the seismic traces that were inaccurately predicted first arrival time. Next, an algorithm is proposed to pick the first arrival time for those seismic traces that have inaccurate picking of the first arrival time predicted by the U-Net deep learning architecture. Seismic horizon interpretation is one of the two main tasks in seismic structural interpretation. An experienced interpreter usually needs weeks or even months to interpret the horizons of a 3D seismic survey. The second part of my dissertation concentrates on accelerating the process of seismic horizon interpretation. Instead of adopting algorithms to automatically extract horizons, I propose to manually interpret horizons on a few key seismic sections and extract horizons under the constraints of manual interpretation.The third part of my dissertation focuses on seismic impedance inversion using physics-informed neural networks. The inputs of the neural networks are seismic amplitude, wavelet, and a low frequency model built using well logs. The output of the neural network is the seismic impedance that corresponds to the seismic trace. The optimization process is minimizing the difference between seismic trace and synthetic seismic waveforms computed from the output impedance log and input wavelet."],"dc:format.medium":["electronic"],"dc:format.mimetype":["application/pdf"],"dc:identifier.other":["1085958"],"dc:identifier.uri":["https://ir.ua.edu/handle/123456789/14409"],"dc:language":["English"],"dc:language.iso":["en_US"],"dc:publisher":["University of Alabama Libraries"],"dc:rights":["All rights reserved by the author unless otherwise indicated."],"dc:title":["Seismic Data Processing and Interpretation via Deep Learning"],"dc:type":["thesis","text"]},"updated_at":"2026-07-27T18:44:20Z"}