Download the full-sized PDF of Gradient projection methods with applications to simultaneous source seismic data processingDownload the full-sized PDF



Permanent link (DOI):


Export to: EndNote  |  Zotero  |  Mendeley


This file is in the following communities:

Graduate Studies and Research, Faculty of


This file is in the following collections:

Theses and Dissertations

Gradient projection methods with applications to simultaneous source seismic data processing Open Access


Other title
Seismic data reconstruction
Least-squares migration
Matrix rank reduction
Random projection
Gradient projection
Simultaneous source
Type of item
Degree grantor
University of Alberta
Author or creator
Cheng, Jinkun
Supervisor and department
Sacchi, Mauricio (Physics)
Examining committee member and department
Poutkaradze, Vakhtang (Math)
Robertsson, Johan (Earth Sciences, ETH Zurich)
Schmitt, Doug (Physics)
Meldrum, Al (Physics)
Heimpel, Moritz (Physics)
Department of Physics
Date accepted
Graduation date
2017-11:Fall 2017
Doctor of Philosophy
Degree level
Simultaneous source acquisition, or blended acquisition, has become an important strategy to reduce the cost of seismic surveys by allowing overlapping between different sources. The major technical challenge associated with this acquisition design is the strong interferences caused by the closely fired shots. This thesis focuses on the separation or the deblending of simultaneous source data via constrained inversion methods. The cost function is the misfit between the predicted and the observed blended data. The constraint is that the desired signal is coherent in the common receiver, common offset, and common midpoint domains when the fire time delay corresponding to each shot is corrected. The simultaneous source interferences would appear incoherent when the randomized source fire scheme is applied. In this thesis, I assume that the desired coherent signal can be represented via a low-rank matrix. The randomly distributed interferences would increase the rank. The coherence constraint for deblending can be implemented effectively by a low-rank constraint in the corresponding data domain. I adopt the gradient projection method that iteratively solves this low-rank constrained inverse problem for deblending. The projection filters are the f-x-y eigenimage filter (Chapter 3) and the Singular Spectrum Analysis filter (Chapter 4) that suppress the source cross-talk artifacts while preserving the unblended signal. Fast implementations of the two reduced-rank filters are achieved via randomized rank-reduction methods. The gradient projection framework is then extended to the direct imaging of simultaneous data via shot-profile least-squares migration.
This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for the purpose of private, scholarly or scientific research. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.
Citation for previous publication

File Details

Date Uploaded
Date Modified
Audit Status
Audits have not yet been run on this file.
File format: pdf (Portable Document Format)
Mime type: application/pdf
File size: 25196691
Last modified: 2017:11:08 17:28:02-07:00
Filename: Cheng_Jinkun_M_201709_PhD.pdf
Original checksum: 1fb66edb8852913d542c0a3a7f638687
Well formed: false
Valid: false
Status message: Invalid page tree node offset=15432139
Status message: Unexpected error in findFonts java.lang.ClassCastException: edu.harvard.hul.ois.jhove.module.pdf.PdfSimpleObject cannot be cast to edu.harvard.hul.ois.jhove.module.pdf.PdfDictionary offset=5885
Status message: Invalid name tree offset=25176358
Status message: Invalid name tree offset=25168925
Status message: Invalid name tree offset=25168925
Activity of users you follow
User Activity Date