The HPE Data Science Institute in partnership with the Geophysical Society of Houston delivered a webinar regarding, “Machine-Learning-Based Log Prediction and its Application to Prestack Depth Migration.”
Machine-learning has impacted multiple industries especially the geoscience field. The geoscience field incorporates high-level data which can be utilized in many spaces.
Mike Perz, MSc, Director of Technology and Innovation, multi-client, onshore at TGS, explored the possibilities of ML-predicted sonic data being used to improve prestack depth migration velocity model building. With 20+ years of experience, Perz has a variety of expertise in the field of seismic technology.
He shared details regarding ARLAS sonic data and if it can improve a velocity model building in a prestack depth migration. He displayed 3D seismic models with and without ARLAS to compare its effects. Perz discussed distinctive challenges within the ARLAS sonic dataset, such as missing segments. In the models, the data and effects vary, especially in a non-ARLAS velocity cube map.
Perz hopes to encourage researchers and everyone else to learn the different aspects of seismic technology and explore how it can deliver better value.