Surface-Related and Internal Multiple Elimination Using Deep Learning

Multiple elimination has always been a key, challenge, and hotspot in the field of hydrocarbon exploration. However, each multiple method comes with one or more limitations at present. The efficiency success approach strongly depend on their corresponding prior assumptions, particular for seismic data acquired from complex geological regions. using deep learning encodes input to levels abstraction decodes those reconstruct primaries without multiples. In this study, we employ classic convolution neural network (CNN) U-shaped architecture which uses extremely few end-to-end training, increasing speed. Then, apply trained predict all data, solves problem difficult global multiples, avoids regularization reduces massive amounts calculation traditional methods. Several synthetic experiments are conducted validate advantages model. results indicate that model powerful generalization ability high removing surface-related multiples internal effectively.