Tim T.Y. Lin

Compressive simultaneous full-waveform simulation

The fact that the computational complexity of wavefield simulation is proportional to the size of the discretized model and acquisition geometry and not to the complexity of the simulated wavefield is a major impediment within seismic imaging. By turning simulation into a compressive sensing problem, where simulated data are recovered from a relatively small number of independent simultaneous sources, we remove this impediment by showing that compressively sampling a simulation is equivalent to compressively sampling the sources, followed by solving a reduced system. As in compressive sensing, this reduces sampling rate and hence simulation costs. We demonstrate this principle for the time-harmonic Helmholtz solver. The solution is computed by inverting the reduced system, followed by recovering the full wavefield with a program that promotes sparsity. Depending on the wavefields sparsity, this approach can lead to significant cost reductions, particularly when combined with the implicit preconditioned Helmholtz solver, which is known to converge even for decreasing mesh sizes and increasing angular frequencies. These properties make our scheme a viable alternative to explicit time-domain finite differences.

Compressed wavefield extrapolation

An explicit algorithm for the extrapolation of one-way wavefields is proposed that combines recent developments in information theory and theoretical signal processing with the physics of wave propagation. Because of excessive memory requirements, explicit formulations for wave propagation have proven to be a challenge in 3D. By using ideas from compressed sensing, we are able to formulate the (inverse) wavefield extrapolation problem on small subsets of the data volume, thereby reducing the size of the operators. Compressed sensing entails a new paradigm for signal recovery that provides conditions under which signals can be recovered from incomplete samplings by nonlinear recovery methods that promote sparsity of the to-be-recovered signal. According to this theory, signals can be successfully recovered when the measurement basis is incoherent with the representation in which the wavefield is sparse. In this new approach, the eigenfunctions of the Helmholtz operator are recognized as a basis that is incoherent with curvelets that are known to compress seismic wavefields. By casting the wavefield extrapolation problem in this framework, wavefields can be successfully extrapolated in the modal domain, despite evanescent wave modes. The degree to which the wavefield can be recovered depends on the number of missing (evanescent) wavemodes and on the complexity of the wavefield. A proof of principle for the compressed sensing method is given for inverse wavefield extrapolation in 2D, together with a pathway to 3D during which the multiscale and multiangular properties of curvelets, in relation to the Helmholz operator, are exploited. The results show that our method is stable, has reduced dip limitations, and handles evanescent waves in inverse extrapolation.

Designing Simultaneous Acquisitions with Compressive Sensing

The goal of this paper is in designing a functional simultaneous acquisition scheme by applying the principles of compressive sensing. By framing the acquisition in a compressive sensing setting we immediately gain insight into not only how to choose the

Unified Compressive Sensing Framework For Simultaneous Acquisition With Primary Estimation

The central promise of simultaneous acquisition is a vastly improved crew efficiency during acquisition at the cost of additional post-processing to obtain conventional source-separated data volumes. Using recent theories from the field of compressive sensing, we present a way to systematically model the effects of simultaneous acquisition. Our formulation form a new framework in the study of acquisition design and naturally leads to an inversion-based approach for the separation of shot records. Furthermore, we show how other inversion-based methods, such as a recently proposed method from van Groenestijn and Verschuur (2009) for primary estimation, can be processed together with the demultiplexing problem to achieve a better result compared to a separate treatment of these problems.

Sparsity-promoting recovery from simultaneous data: a compressive sensing approach

Summary Seismic data acquisition forms one of the main bottlenecks in seismic imaging and inversion. The high cost of acquisition work and collection of massive data volumes compel the adoption of simultaneous-source seismic data acquisition - an emerging technology that is developing rapidly, stimulating both geophysical research and commercial efforts. Aimed at improving the performance of marine- and land-acquisition crews, simultaneous acquisition calls for development of a new set of design principles and post-processing tools. Leveraging developments from the field of compressive sensing the focus here is on simultaneous-acquisition design and sequential-source data recovery. Apart from proper compressive sensing sampling schemes, the recovery from simultaneous simulations depends on a sparsifying transform that compresses seismic data, is fast, and reasonably incoherent with the compressive-sampling matrix. Using the curvelet transform, in which seismic data can be represented parsimoniously, the recovery of the sequential-source data volumes is achieved using the sparsity-promoting program — SPGL1, a solver based on projected spectral gradients. The main outcome of this approach is a new technology where acquisition related costs are no longer determined by the stringent Nyquist sampling criteria.

Estimating Primaries by Sparse Inversion in a Curvelet-like Representation Domain

We present an uplift in the fidelity and wavefront continuity of results obtained from the Estimation of Primaries by Sparse Inversion (EPSI) program by reconstructing the primary events in a hybrid wavelet-curvelet representation domain. EPSI is a multiple removal technique that belongs to the class of wavefield inversion methods, as an alternative to the traditional adaptive-subtraction process. The main assumption is that the correct primary events should be as sparsely-populated in time as possible. A convex reformulation of the original EPSI algorithm allows its convergence property to be preserved even when the solution wavefield is not formed in the physical domain. Since wavefronts and edge-type singularities are sparsely represented in the curvelet domain, sparse solutions formed in this domain will exhibit vastly improved continuity when compared to those formed in the physical domain, especially for the low-energy events at later arrival times. Furthermore, a wavelet-type representation domain will preserve sparsity in the reflected events even if they originate from non-zero-order discontinuities in the subsurface, providing an additional level of robustness. This method does not require any changes in the underlying computational algorithm and does not explicitly impose continuity constraints on each update.

Interpolating solutions of the helmholtz equation with compressed sensing

We present an algorithm which allows us to model wavefields with frequency-domain methods using a much smaller number of frequencies than that typically required by the classical sampling theory in order to obtain an alias-free result. The foundation of the algorithm is the recent results on the compressed sensing, which state that data can be successfully recovered from an incomplete measurement if the data is sufficiently sparse. Results from numerical experiment show that only 30% of the total frequency spectrum is need to capture the full wavefield information when working in the hard 2D synthetic Marmousi model.

Sparsity-promoting Migration with Surface-related Multiples

Multiples, especially the surface-related multiples, form a significant part of the total up-going wavefield. If not properly dealt with, they can lead to false reflectors in the final image. So conventionally practitioners remove them prior to migration. Recently research has revealed that multiples can actually provide extra illumination so different methods are proposed to address the issue that how to use multiples in seismic imaging, but with various kinds of limitations. In this abstract, we combine primary estimation and sparsity-promoting migration into one convex-optimization process to include information from multiples. Synthetic examples show that multiples do make active contributions to seismic migration. Also by this combination, we can benefit from better recoveries of the Greens function by using sparsity-promoting algorithms since reflectivity is sparser than the Greens function.

Compressive sensing applied to full-wave form inversion

Full-waveform inversions high demand on computational resources forms, along with the non-uniqueness problem, the major impediment withstanding its widespread use on industrial-size datasets. Turning modeling and inversion into a compressive sensing prob

Migration with surface‐related multiples from incomplete seismic data

Seismic acquisition is confined by limited aperture that leads to finite illumination, which, together with other factors, hinders imaging of subsurface objects in complex geological settings such as salt structures. Conventional processing, including surface-related multiple elimination, further reduces the amount of information we can get from seismic data. With the growing consensus that multiples carry valuable information that is missing from primaries, we are motivated to exploit the extra illumination provided by multiples to image the subsurface. In earlier research, we proposed such a method by combining primary estimation and sparsity-promoting migration to invert for model perturbations directly from the total up-going wavefield. In this abstract, we focus on a particular case. By exploiting the extra illumination from surface-related multiples, we mitigate the effects caused by migrating from incomplete data with missing sources and missing near-offsets.