Abstract:Multiples are usually regarded as noise in conventional seismic data processing. However, multiples are also real reflections from structural interfaces in the subsurface. Compared with primaries, multiples usually provide more balanced illumination and contain more structural information because of the smaller reflection angles and longer wavepaths. Instead of multiple suppression, multiple imaging has attracted increasingly more attention in recent years. The most commonly used migration method for multiples is performed by replacing the source wavelet with recorded data and using separated multiples as the receiver record. Then, the image of the multiples is obtained by the application of the crosscorrelation imaging condition, which is widely used in conventional migration. However, during the imaging procedure, events are matched based on their propagation times only. Crosscorrelation of unrelated events leads to heavy crosstalk, and image artifacts are introduced in the image of multiples. To overcome this shortcoming, we have introduced the stereographic imaging condition for the one-way wave-equation migration of multiples. By adding a local-slope constraint (the local slope of the extrapolated wavefields at every position and time), the stereographic imaging condition takes the local spatial coherence of the extrapolated wavefields into account. Events can be matched based not only on the propagating times but also the local slopes. Therefore, crosstalk artifacts caused by the interference of unrelated events can be efficiently suppressed. Furthermore, to improve the computational accuracy and efficiency of our approach, plane-wave destructors are introduced to estimate the reflector slopes. In this way, the need for excessive loops over the local slopes in the x-z domain during application of the stereographic imaging condition can be avoided by selecting the local slopes in a proper range. Better migration results of multiples are obtained in numerical tests, which verify the feasibility and effectiveness of our approach.

Enhancing One-Way Wave Equation-Based Migration with Deep Learning

Steeply-dipping structures migration based on the wavefield vertical-and horizontal-extrapolation

One‐Way Wave Equation Least‐Squares Migration Based on Illumination Compensation

Deep Learning for Enhancing Multi-Source Reverse Time Migration

Elastic Reverse Time Migration Without Wavefield Decomposition

Deep-learning for Accelerating Prestack Correlative Least-Squares Reverse Time Migration

Imaging Of Radar And Shallow Seismic Data Using Full Wave Equation Migration In The Wavelet Domain

Deep learning-based point spread function deconvolution for migration image deblurring

One-step Sparse-Matrix Method for Full-Wave-equation Depth Migration

GAN-enhanced directional seismic wavefield decomposition and its application in reverse-time migration

Improving the image quality of elastic reverse-time migration in the dip-angle domain using deep learning

Deep Convolutional Neural Network and Sparse Least Squares Migration

Mesh-free Radial-Basis-function-generated Finite Differences and Their Application in Reverse Time Migration

An Effective Denoising Strategy for Wave Equation Migration Based on Propagation Angles

A Multi-channel Input Framework for Structure Displacement Response Prediction Using Convolutional Neural Network

Least-squares reverse-time migration of simultaneous source with deep-learning-based denoising

Generative Adversarial Network-Enhanced Directional Seismic Wavefield Decomposition and Its Application in Reverse Time Migration

Synthesized plane wave preserved amplitude prestack depth migration based one-way wave equation

High-SNR Staining Algorithm for One-Way Wave Equation-Based Modelling and Imaging

One-way wave-equation migration of multiples based on stereographic imaging condition

Deep-Learning Full-Waveform Inversion Using Seismic Migration Images