Local Linear Events Extraction and Filtering in the Presence of Time-shifts

Static shifts are detrimental for any algorithm that is based on the lateral coherency of the seismic events. We present a multidimensional extraction of linear events, where the time shifts are estimated alongside the other attributes of the extracted event (slopes, and wavelets), leading to a much more effective extraction. Different time-shifts are estimated for each extracted event, making the algorithm more generic than an approach based on classical static shifts. We demonstrate the effectiveness of this algorithm on synthetic and field data.