Seismic Inversion

Seismic inversion aims to reconstruct subsurface geological structures from seismic wave recordings, a challenging inverse problem due to data limitations and noise. Current research heavily utilizes deep learning, employing convolutional neural networks (CNNs), encoder-decoder architectures, and physics-informed neural networks (PINNs) to improve inversion accuracy and speed, often incorporating techniques like compressed sensing and joint inversion of multiple data types (e.g., gravity and seismic). These advancements are crucial for applications such as oil and gas exploration, CO2 sequestration monitoring, and geotechnical assessments, enabling more efficient and reliable subsurface imaging.