Self-supervised learning for anti-aliasing seismic data interpolation
IMAGE 2021
Pengyu Yuan, Shirui Wang, Wenyi Hu, Prashanth Nadukandi, German Ocampo Botero, Xuqing Wu, Jiefu Chen, and Hien Van Nguyen
Abstract
Reconstruction of seismic data in exploration geophysics is an important but challenging task in seismic data processing. Different machine-learning based approaches have been de- veloped to solve this ill-posed problem and achieved great progress. However, most of the available machine-learning tools use supervised learning method where a good training dataset with many complete shot–gathers are required to train the model. Existing unsupervised learning approach based on a generative model is able to reconstruct the general image pat- terns in a shot–gather but fails to accurately resolve the fine features, especially when aliasing is a main concern. In this work, we propose a self-supervised learning method that com- bines a blind-trace network and an automatic spectrum sup- pression technique for high-accuracy seismic data reconstruc- tion. We present some examples covering challenging scenar- ios, e.g. a wide range of randomly or regularly decimated seis- mic data, where our method consistently performs accurate re- constructions.