Paper submitted to SEG Transforming Energy Exploration: Technology, Sustainability, and Innovation Symposium, by G. Stock, D. Brookes, O. Brusova, A. Knowles, S. Mannick, M. Roberts, R. Yates J. Kumar and P. Lee (TGS).
Abstract
Reprocessing multi-vintage marine seismic data to a consistent broadband standard remains a major challenge, particularly when acquisition parameters vary significantly between surveys and auxiliary acquisition information is incomplete or unavailable. Deghosting is a critical component of modern marine seismic reprocessing, enabling recovery of low- and high-frequency bandwidth, improved amplitude stability, and enhanced temporal resolution through the removal of surfacerelated ghost reflections. This abstract presents an augmented machine-learning (ML) deghosting workflow applied to a large multi-survey reprocessing project in the Krishna–Godavari (KG) Basin offshore India. The approach combines a generalized ML deghosting model with survey-specific adaptation derived from limited deterministic deghosting results. The workflow delivers stable broadband spectra, improved seismic imaging, and consistent results across surveys with varying tow depths and acquisition geometries, while maintaining efficient turnaround times.