There has been growing interest in the use of machine learning technologies for processing and interpreting seismic data. Many procedures that traditionally have been performed using deterministic methods and algorithms can be effectively replaced by neural networks and other artificial intelligence methodologies, improving simplicity, efficiency and automation.

This paper discusses the generation of synthetic 3D seismic data for training neural networks to solve a variety of seismic processing, interpretation and inversion tasks. It presents a methodology for generating on-the-fly simulated post-migration (1D) synthetic data in 3D, which are high resolution and look similar to real data.