A comparative study of data- and image- domain LSRTM under velocity-impedance parametrization

TL;DR

This study compares data-domain and image-domain LSRTM under a velocity–impedance parametricization with logarithmic scaling, implemented in SMIwiz. It demonstrates a first 3D data-domain multiparameter LSRTM and analyzes PSF-based image-domain alternatives, revealing that data-domain LSRTM yields superior reflectivity but at higher computational cost, while image-domain methods struggle with multiparameter coupling yet remain effective for impedance-only inversion. The findings are validated on 2D Marmousi, 3D Overthrust, and Viking Graben datasets, showing consistent convergence advantages for preconditioned CGNR and outlining practical trade-offs between accuracy, memory, and efficiency. The work lays a pathway for extensions to elastic media and more advanced Hessian treatments, with open-source code available in SMIwiz.

Abstract

Least-squares reverse time migration (LSRTM) is one of the classic seismic imaging methods to reconstruct model perturbations within a known reference medium. It can be computed in either data or image domain using different methods by solving a linear inverse problem, whereas a careful comparison analysis of them is lacking in the literature. In this article, we present a comparative study for multiparameter LSRTM in data- and image- domain in the framework of SMIwiz open software. Different from conventional LSRTM for recovering only velocity perturbation with variable density, we focus on simultaneous reconstruction of velocity and impedance perturbations after logorithmic scaling, using the first-order velocity-pressure formulation of acoustic wave equation. The first 3D data-domain LSRTM example has been performed to validate our implementation, involving expensive repetition of Born modelling and migration over a number of iterations. As a more cost-effective alternative, the image-domain LSRTM is implemented using point spread function (PSF) and nonstationary deblurring filter. Dramatic disctinctions between data and image domain methods are discovered with 2D Marmousi test: (1) The data-domain multiparameter inversion provides much better reconstruction of reflectivity images than image-domain approaches, thanks to the complete use of Hessian in Krylov space; (2) The poor multiparameter image-domain inversion highlights the limitation of incomplete Hessian sampling and strong parameter crosstalks, making it difficult to work in practice; (3) In contrast, monoparameter image-domain inversion for seismic impedance is found to work well. These observations have been further validated on Viking Graben Line 12 dataset.

Figures (22)

• Figure 1: The true Marmousi model (1st row), smoothed one (2nd row) associated with velocity (1st column), density (2nd column) and impedance (3rd column)
• Figure 2: (a) the first shot modelled from true Marmousi model; (b) muting mask; (c) simulated data from smooth Marmousi model; (d) the residual data as the input for linearized waveform inversion after muting direct arrivals and far-offset diving waves.
• Figure 3: RTM images for logarithmic parameters: (a) wavespeed $\ln V_p$ and (b) impedance $\ln I_p$
• Figure 4: Normalized misfit of data-domain LSRTM with (PCGNR) and without (CGNR) preconditioning
• Figure 5: (a, b) true $\delta \ln V_p$ and $\delta\ln I_p$; (c, d) $\delta \ln V_p$ and $\delta \ln I_p$ inverted by CGNR; (e, f) $\delta \ln V_p$ and $\delta \ln I_p$ inverted by PCGNR.