CO2Seg: Automatic CO2 Segmentation From 4-D Seismic Image Using Convolutional Vision Transformer

To tackle the pressing issue of climate change stemming from carbon emissions, carbon capture and storage (CCS) projects have emerged worldwide, which aim to store carbon dioxide (CO2) produced during industrial production in subsurface geological structures. To ensure the efficacy of these projects, 4-D seismic surveys are conducted to monitor the stored CO2 and identify potential leakage at an early stage. In recent years, deep learning (DL) has been widely employed for seismic data interpretation, which has shown promising results in terms of objectivity and efficiency when compared to manual interpretation. In this study, we address the CO2 monitoring challenge using a 3-D encoder–decoder network with convolutional vision transformer (CvT) called CvTNet, through the supervised learning scheme. By formulating the CO2 monitoring task as an image segmentation problem, we use CvTNet to generate a 3-D CO2 probability image from a 4-D seismic image. CvTNet leverages the CvT module, which provides superior dynamic attention and global context compared to convolutional neural networks. We evaluate the effectiveness of CvTNet on the Sleipner CCS project, using a 4-D seismic image (comprising a 3-D baseline image from 1994 and a 3-D time-lapse image from 2010) and a CO2 probability image from 2010 as the CvTNet training input and label, respectively. We apply the trained model to seismic images from other monitoring years to analyze CO2 plume growth during the Sleipner CCS project. Tests indicate that CvTNet achieves higher CO2 segmentation accuracy than U-Net and can be generalized across other 4-D seismic images.