Impedance Parameterization-Based Time-Lapse Least-Squares Reverse Time Migration with Structure-Guided Constraints

Abstract Time-lapse seismic imaging acts as an essential tool to monitor reservoir changes. To accurately capture these subsurface variations, time-lapse seismic imaging usually must overcome intense imaging artifacts and noise. These artifacts easily contaminate the time-lapse difference sections, severely obscuring genuine reservoir changes and leading to potential misinterpretations of dynamic change. To address this limitation, we propose an integrated time-lapse least-squares reverse time migration (LSRTM) inversion framework that combines acoustic impedance parameterization with structure-guided constraints. Unlike conventional velocity-parameterized least-squares reverse time migration, we directly select impedance as the perturbation parameter and invert for the corresponding reflectivity model. This parameterization method preliminarily suppresses low-wavenumber background noise, producing clear imaging results without the artifacts commonly observed in conventional velocity-parameterized LSRTM. Next, we use these imaging results to extract accurate structural information. By incorporating this information into the time-lapse inversion as structure-guided constraints, we force the gradient updates to follow the true geological structures. This structure-guided constraint method further eliminates the residual migration artifacts in the imaging results. Tests on modified Marmousi and BP gas chimney synthetic models demonstrate that our method achieves high-quality, artifact-free time-lapse difference images with significantly fewer iterations than conventional least-squares reverse time migration. The framework offers a practical solution for reliable reservoir characterization in structurally complex settings, particularly where conventional methods fail due to strong noise contamination.