Adaptive image focusing

Seismic image quality dependency on the migration velocity model becomes progressively higher as geological complexity increases. Advancement in more theoretically rigorous model estimation technologies such as waveform inversion will gradually improve our model building capabilities over time to meet this challenge, but it is also important to realize that certain degrees of imperfection in the velocity models will always be expected, at least in the foreseeable future. There is therefore a need to adjust/improve images as part of the migration process to account for this imperfection, and furthermore to mitigate the inadequacy of the physics we use at various stages in the data processing pipeline. To this end, we look into adaptive image focusing, an automated workflow aiming at extracting improved imaging results from imperfect migration velocity models, which may involve amplitude balancing and alignment (to directly construct improved gathers) and/or noise suppression (to remove destructive energy from the stack). As a first attempt to achieve such a workflow, we emphasize the noise suppression aspect of the image focusing problem and demonstrate in this work the possibility to identify and suppress non-contributing shot-record migration results in the time-lag extended image domain. We explain the basic processing steps with a synthetic example which shows promising results.