Abstract:Time-lapse (or 4-D) seismic data play an important role in monitoring the spatial CO 2 distribution during and after the injection period. However, traditional interpretation or prediction of CO 2 distribution is time-consuming and might be sensitive to the quality of 4D seismic data. To solve these problems, we propose a deep-learning-based method to efficiently and accurately characterize CO 2 plumes in time-lapse seismic data. We first introduce a workflow to build 3-D realistic impedance models containing CO 2 plumes with various shapes, sizes, and locations. From the impedance models, we then simulate synthetic seismic datasets and automatically obtain the corresponding CO 2 label volumes. We extract real noise from field seismic datasets and add the noise to the synthetic ones to make them more realistic. We further construct a diverse and realistic training dataset with the combination of synthetic data containing CO 2 plumes and real data without CO 2 plumes that are randomly cropped from field seismic data before CO 2 injection. We finally utilize the training datasets without any human labeling to train a 3-D deep U-shape convolutional neural network for detecting CO 2 plumes in the Sleipner time-lapse seismic images. Compared to traditional interpretation methods that take several days or even weeks, our method takes only 29 s using one graphics processing unit (GPU) to predict CO 2 plumes in a 512*512*256 seismic volume. Besides, our CO 2 prediction can achieve 95.8% accuracy (compared to the manual interpretation) and could distinguish reflections of CO 2 plumes from the ones of pre-existing fluids, thin layers, and noise. To more accurately characterize the CO 2 plumes migration, we use dynamic image warping to compute relative shifts that register the time-lapse seismic volumes before and after CO 2 injection and then apply the same shifts to the predicted CO 2 plumes. By doing this, we are able to reduce the inconsistencies that may be introduced by acquisition, processing, push-down effect (velocity decrease by injected CO 2 ), and pull-up effect (wavelet distortion), which is helpful to more accurately characterize the CO 2 plumes migration.

A Real-time Identification Method Based on Static-dynamic Optical Flow Features for Deep-sea Hydrothermal Plumes

Numerical Modeling of Hydrodynamic Processes of Deep-Sea Hydrothermal Plumes: A Case Study on Daxi Hydrothermal Field, Carlsberg Ridge

A New Flow Pattern Identification Method for Gas–Liquid Two-Phase Flow in Small Channel Based on an Improved Optical Flow Algorithm

Gas-liquid Two-Phase Flow Pattern Identification Based on Image Processing

DeepPTV: Particle Tracking Velocimetry for Complex Flow Motion via Deep Neural Networks

Visual identification of oscillatory two-phase flow with complex flow patterns

Flow Regime Identification of Mini-Pipe Gas-Liquid Two-Phase Flow Based on Textural Feature Series

Real-time Detection of Deep-Sea Hydrothermal Plume Based on Machine Vision and Deep Learning

Passive Acoustic Identification of Bubble Flow Regime Based on S Ynchrosqueezing Wavelet Transform and Deep Learning

A Plankton Detection Method Based on Neural Networks and Digital Holographic Imaging

Visual Detection And Feature Recognition Of Underwater Target Using A Novel Model-Based Method

Deep learning for characterizing CO 2 migration in time-lapse seismic images

Automatic Identification and Suppression of Random Noise and Methods for Profile Splicing in the Sub-Bottom Profile of Deep Water

Oil film identification via windspeed interference estimation using hyperspectral data

Underwater Gas Leakage Flow Detection and Classification Based on Multibeam Forward-Looking Sonar

Adaptive river flow measurement method based on spatiotemporal image velocimetry and optical flow

SeisLFMFlow: Seismic Common Image Gathers Enhancement Using Self-Supervised Optical Flow Estimation Based on Local Feature Matching

Ultraviolet-induced fluorescence of oil spill recognition using a semi-supervised algorithm based on thickness and mixing proportion–emission matrices

Modelling and detection of submarine bubble plumes using seismic oceanography

Dynamic Texture Recognition by Spatio-Temporal Multiresolution Histograms

A Noncontact Technique for Wave Measurement Based on Thermal Stereography and Deep Learning