Zero-offset data estimation using CNN for applying 1D full waveform inversion

Abstract Full waveform inversion (FWI) in the time domain has limitations due to large computing and memory requirements. Some studies have addressed this problem by using machine learning techniques. Most FWI directly estimate subsurface velocity structure training seismic data generated through various synthetic models obtain structure. In study, we propose a method convert common midpoint (CMP) gather zero-offset at CMP location convolutional neural network (CNN) increase efficiency for FWI. As data, use exploration geometry source signature of field data. Since proposed performs converted it can be performed more efficiently than existing multichannel However, is difficult apply that not been used training. To verify method, applied model as well focused It confirmed proper was obtained.