Seismic wave velocity modeling method based on SR Transform and multi-receptive-field module

The invention discloses a seismic wave velocity modeling method based on an SR Transform and a multi-receptive field module network. A multi-receptive-field module is introduced to excavate correlation of seismic wave frequencies, and new seismic wave features are constructed based on the learned correlation, so that high-quality seismic wave speed model construction is realized. The SR Transform Block and multi-receptive field module-based network proposed by the invention for the first time can effectively help geological exploration to solve the seismic velocity problem, so that the model can capture the long-distance dependency relationship between features, and gradient disappearance is avoided. A large number of experiments carried out on SEG Salt and Sulled data sets show that the model obtains a good result in the aspect of speed modeling. In future work, we will continue to explore how to better learn the semantics of shot records and effectively integrate them for velocity modeling. 本发明公开了基于SR Trans