A5071027426- Seismic Imaging and Inversion Techniques
- Seismic Waves and Analysis
- Reservoir Engineering and Simulation Methods
- Drilling and Well Engineering
- Hydraulic Fracturing and Reservoir Analysis
- Geophysical Methods and Applications
- Ultrasound Imaging and Elastography
- Photoacoustic and Ultrasonic Imaging
- Image and Signal Denoising Methods
- Ultrasonics and Acoustic Wave Propagation
- Advanced MRI Techniques and Applications
- Medical Imaging Techniques and Applications
- Ultrasound and Hyperthermia Applications
- Seismology and Earthquake Studies
- Advanced Computational Techniques in Science and Engineering
- Fault Detection and Control Systems
- Marine and coastal plant biology
- Medieval European Literature and History
- Radiomics and Machine Learning in Medical Imaging
- Comparative Literary Analysis and Criticism
- Generative Adversarial Networks and Image Synthesis
- Cell Image Analysis Techniques
- Coastal wetland ecosystem dynamics
- Gaussian Processes and Bayesian Inference
- Microwave Imaging and Scattering Analysis
Imperial College London
2013-2024
Geociencias Barcelona
2008
We have developed and implemented a robust practical scheme for anisotropic 3D acoustic full-waveform inversion (FWI). demonstrate this on field data set, applying it to 4C ocean-bottom survey over the Tommeliten Alpha in North Sea. This shallow-water set provides good azimuthal coverage offsets of 7 km, with reduced maximum offset about 11 km. The reservoir lies at crest high-velocity antiformal chalk section, overlain by 3000 m clastics within which low-velocity gas cloud produces seismic...
Conventional full-waveform seismic inversion attempts to find a model of the subsurface that is able predict observed waveforms exactly; it proceeds by minimizing difference between and predicted data directly, iterating in series linearized steps from an assumed starting model. If this too far removed true model, then approach leads spurious which are cycle skipped with respect data. Adaptive waveform (AWI) provides new form (FWI) appears be immune problems otherwise generated skipping. In...
Magnetic resonance imaging and X-ray computed tomography provide the two principal methods available for brain at high spatial resolution, but these are not easily portable cannot be applied safely to all patients. Ultrasound is universally safe, existing modalities image usefully inside adult human skull. We use in silico simulations demonstrate that full-waveform inversion, a computational technique originally developed geophysics, able generate accurate three-dimensional images of with...
We propose an extended full-waveform inversion formulation that includes general convex constraints on the model. Though full problem is highly nonconvex, overarching optimization scheme arrives at geologically plausible results by solving a sequence of relaxed and warm-started constrained subproblems. The combination box, total variation, successively asymmetric variation allows us to steer free from parasitic local minima while keeping estimated physical parameters laterally continuous in...
Computational models of acoustic wave propagation are frequently used in transcranial ultrasound therapy, for example, to calculate the intracranial pressure field or phase delays correct skull distortions. To allow intercomparison between different modeling tools and techniques by community, an international working group was convened formulate a set numerical benchmarks. Here, these benchmarks presented, along with results. Nine increasing geometric complexity defined. These include...
Posidonia oceanica is a widespread coastal Mediterranean seagrass which accumulates in its subsurface large quantities of organic material derived from roots, rhizomes and leaf sheaths embedded sandy sediments. These deposits may be up to several meters thick as they accumulate over thousands years forming the matte , whose high content carbon plays major role global ocean cycle. In this study, very high‐resolution seismo‐acoustic methods were applied image features P. meadow at Portlligat...
Summary We present a new method for performing full-waveform inversion that appears to be immune the effects of cycle skipping – Adaptive Waveform Inversion (AWI). The uses Wiener filters match observed and predicted data. is formulated so model updated in direction drives these towards delta functions at zero lag, which point true has been recovered. computationally efficient, it universally applicable, recovers correct when conventional FWI fails entirely.
Summary Conventional FWI minimises the direct differences between observed and predicted seismic datasets. Because data are oscillatory, this approach will suffer from detrimental effects of cycle skipping if starting model is inaccurate. We reformulate so that it instead adapts to using Wiener filters, then iterates improve by forcing filters towards zero-lag delta functions. This adaptive scheme demonstrated on synthetic where shown be immune skipping, able invert successfully for which...
Given appropriate data acquisition, processing to remove nonprimary arrivals, and use of an accurate migration algorithm, it is the quality subsurface velocity model that typically controls imaging can be obtained from salt-affected seismic data. Full-waveform inversion has potential improve accuracy, resolution, repeatability, speed with which such models generated, but, in absence starting model, difficult realize practice. Presented are successful results, synthetic subsalt models, using...
Adaptive waveform inversion (AWI) reformulates the misfit function used to perform full-waveform (FWI), so that it no longer contains local minima related cycle skipping. It does this by finding a model drives ratio of predicted and observed data sets unity rather than driving difference between these two zero as is case for conventional FWI. We apply AWI 3D field set acquired over pervasive gas cloud in North Sea, comparing its performance with FWI variety circumstances. When starting from...
Advanced ultrasound computed tomography techniques like full-waveform inversion are mathematically complex and orders of magnitude more computationally expensive than conventional imaging methods. This computational algorithmic complexity, a lack open-source libraries in this field, represent barrier preventing the generalised adoption these techniques, slowing pace research, hindering reproducibility. Consequently, we have developed Stride, an Python library for solution large-scale...
Ultrasound breast imaging is a promising alternative to conventional mammography because it does not expose women harmful ionising radiation and can successfully image dense tissue. However, ultrasound only provides morphological information with limited diagnostic value. computed tomography (USCT) uses energy in both transmission reflection when the provide more diagnostically relevant quantitative tissue properties, but often based on time-of-flight or similar ray approximations of wave...
Successful full-waveform inversion (FWI) of 3D seismic data typically requires low-frequency content in the field coupled with an accurate starting velocity model. In these circumstances, two fundamental questions always arise: (1) is model sufficiently given that are available, and (2) will iterate towards global minimum, or it instead become trapped locally leading to erroneous final model? We present a robust objective means answer both questions. The diagnostic feature we use achieve...
We demonstrate the application of elastic 3D full-waveform inversion (FWI) to a field dataset. first analyze and validate its performance using synthetic data generated from version Marmousi model. show that FWI can recover both p-wave shear-wave velocity models only single-component hydrophone data. then apply multi-component OBC air-gun dataset North Sea, it successfully recovers structure shallow channels.
Despite being a low-cost, portable and safe medical imaging technique, transcranial ultrasound is not used widely in adults because of the severe degradation distortion signals caused by skull. Full-waveform inversion (FWI) has recently been found to have potential as an effective method for tomography obtain high-quality, subwavelength-resolution acoustic models brain using low-frequency data. In this study first demonstration recovering high-resolution 2-D reconstruction skull phantom...
By minimizing the difference between synthetic and field data sets, full waveform inversion (FWI) can produce high resolution fidelity earth model parameters that are not resolvable by commonly used ray-based tomography. In this paper, we share our experience on applying 3-D acoustic time-domain to an OBC set collected over Valhall in North Sea.
Full waveform seismic inversion, as currently commercially available in 3D, uses the acoustic approximation to wave equation, and generally ignores effects of elasticity, attenuation anomalous density variations. We examine consequences these practical compromises by inverting 3D synthetic data using different approximations physics propagation forward inverse modelling. also invert portions data, amplitudes. conclude that current FWI practice works well if amplitudes reflected energy are...
Bayesian methods are a popular research direction for inverse problems. There variety of techniques available to solve Bayes' equation, each with their own strengths and limitations. Here, we discuss stochastic variational inference (SVI), which solves equation using gradient-based methods. This is important applications time-limited (e.g. medical tomography) or where solving the forward problem expensive adjoint methods). To evaluate use SVI in both these contexts, apply it ultrasound...
Summary We extend full-waveform inversion by Wavefield Reconstruction Inversion including convex constraints on the model. Contrary to conventional adjoint-state formulations, Wave-field has advantage that Gauss-Newton Hessian is well approximated a diagonal scaling, which allows us add constraints, such as box- and edge-preserving total-variation constraint, square slowness without incurring significant increases in computational costs. As examples demonstrate, these yields far superior...
Summary Adaptive Waveform Inversion (AWI) was introduced by Warner & Guasch (2014) as a means to avoid cycle skipping during full-waveform inversion. Here we demonstrate the robustness of this new method applying it three challenging problems: 3D field dataset without an accurate velocity model start inversion; highly realistic blind synthetic that contains elastic effects, attenuation, unknown density and ambient noise; simple where inversion proceeds using wrong source wavelet. AWI...
Summary We have developed an accurate and robust methodology that is capable of inverting seismic data in salt-affected environments, to obtain a highly resolved velocity model, without the use travel-time tomography or explicit salt flooding, circumstances where conventional full-waveform inversion (FWI) similar approaches fail. A reflection-driven inversion, combination with total variation (TV) constraints applied adaptive waveform (AWI) objective function, provides building blocks...
Summary We present a newly proposed method for performing full-waveform inversion that appears to be immune the effects of cycle skipping – Adaptive Waveform Inversion (AWI). The uses Wiener filters match observed and predicted data. is formulated so model updated in direction drives these towards delta functions at zero lag, which point true has been recovered. computationally efficient, it universally applicable, recovers correct when conventional FWI fails entirely.
Ultrasound computed tomography techniques have the potential to provide clinicians with 3D, quantitative and high-resolution information of both soft hard tissues such as breast or adult human brain. Their practical application requires accurate modelling acquisition setup: spatial location, orientation, impulse response each ultrasound transducer. However, existing calibration methods fail accurately characterise these transducers unless their size can be considered negligible when compared...
We present a new phase-unwrapped full-wavefield inversion (FWI) methodology for applying the technique to seismic data directly from poor or simple starting model in an automated, robust manner. The well-known difficulty that arises with is 'cycle-skipped' relationship between predicted and observed at useable frequencies. local minimum convergence of cycle-skipped one root causes inaccurately recovered models practical applications FWI. Further it why date have focussed on favourable...