Sparse multi-channel simultaneous seismic impedance inversion in the Laplace–Fourier domain for well-free low-frequency recovery
Seismic acoustic impedance (AI) inversion is fundamentally ill-posed due to the band-limited nature of recorded seismic wavefields. In frontier exploration areas lacking reliable well controls, establishing an accurate low-frequency background model (typically 0–10 Hz) is difficult. Furthermore, conventional trace-by-trace inversion schemes neglect lateral spatial correlations, rendering results highly sensitive to random noise and disruptive to stratigraphic continuity. To simultaneously overcome these dual limitations, we propose a novel multi-channel simultaneous inversion framework that integrates deterministic sparse regularization with the Laplace–Fourier domain. Unlike conventional methods that rely on well-log interpolation for low-frequency compensation, our approach estimates low-frequency AI trends from band-limited seismic data through Laplace damping, subject to wavelet calibration and the adopted background-scale constraints. Concurrently, multi-channel structural constraints are imposed via total variation regularization to preserve sharp lithological boundaries and enforce lateral continuity across adjacent traces. The resulting large-scale multi-trace AI optimization problem is solved efficiently using the alternating direction method of multipliers coupled with a two-dimensional Sylvester equation, bypassing the need to store massive sparse coefficient matrices. Comprehensive tests on synthetic data—with explicitly high-pass filtered inputs—and field data demonstrate that the proposed method estimates the 0–10 Hz background trend without a spatially interpolated well-log model under the adopted initialization and calibration assumptions, improves lateral continuity and the delineation of impedance contrasts and stratigraphic boundaries in the tested examples, and exhibits superior noise robustness compared to conventional linear baselines.
Authors
- Guangzhao Guo
- Yuqiang Zhang (ORCID: https://orcid.org/0009-0008-4590-654X)
- Xiaotao Wen
- Kai Lin
- Lian Zhao
Institutions
- Chengdu University of Technology (CN)
- China University of Petroleum, East China (CN)
Publication Details
- Journal
- Journal of seismic exploration
- Published
- 2026-11-20
- DOI
- https://doi.org/10.36922/jse026330152
- Primary Topic
- Seismic Imaging and Inversion Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00