"image-domain LSRTM data"

"RTM images are jointly constrained by limited acquisition aperture, band-limited source wavelets, and errors in the background velocity model. These factors often cause spatially varying resolution loss, uneven illumination, and amplitude distortion. These problems are more obvious in deep structures and poorly illuminated areas. Data-domain least-squares reverse time migration (LSRTM) improves imaging resolution effectively. However, repeated migration and demigration lead to high computational cost and low efficiency. To address this problem, we propose an image-domain learning-based LSRTM framework (DL-ID-LSRTM). The framework introduces degradation information related to migration conditions as conditional constraints during inversion. A forward reconstructability constraint is also introduced to improve physical consistency. Specifically, the proposed method fuses the RTM migration velocity model with point spread function (PSF) description information. These inputs parameterize local degradation features caused by migration. These features depict local focusing behavior and illumination patterns. A spatially adaptive feature modulation mechanism then injects these conditional features into a multiscale inversion backbone. This design guides reflectivity reconstruction. In addition, this paper designs a closed-loop forward reconstruction branch. Under the same migration conditions, the predicted reflectivity is used to synthesize RTM profile. A consistency loss is then imposed in the input domain. This loss keeps the reconstructed result matched to the input RTM image. This design suppresses shortcut learning and reduces nonphysical artifacts. The dataset includes a blind generalization test case. In comparison with RTM, U-Net-based image-domain LSRTM, and data-domain LSRTM with limited iterations, DL-ID-LSRTM yields estimates that align more closely with the actual subsurface reflectivity. The results demonstrate improved structural and deep-event continuity, reduced migration artifacts, and more balanced amplitude responses."