Diffraction separation and imaging using multichannel singular-spectrum analysis

Abstract

Seismic diffractions are the responses of small-scale discontinuous structures. They contain subwavelength geologic information. Thus, diffractions can be used for high-resolution imaging. The energy of diffractions is generally much weaker than that of reflections. Therefore, diffracted energy is typically masked by specular reflected energy. Diffraction/reflection separation is a crucial preprocessing step for diffraction imaging. To resolve the diffraction-separation problem, we have developed a method based on the multichannel singular-spectrum analysis (MSSA) algorithm for diffraction separation by reflection suppression. The MSSA algorithm uses the differences in the kinematic and dynamic properties between reflections and diffractions to suppress time-linear signals (reflections) and separate weaker time-nonlinear signals (diffractions) in the common-offset or poststack domain. For the time-linear signals, the magnitudes of the singular values are proportional to the energy strength of the signals. The stronger the energy of a component of the linear signals is, the larger the corresponding singular values will be. The singular values of reflections and diffractions have dissimilar spatial distributions in the singular-value spectrum because of the differences in their linear properties and energy. Only the singular values representing diffractions are selected to reconstruct seismic signals. Synthetic data and field data are used to test our method. The results reveal the good performance of the MSSA algorithm in enhancing diffractions and suppressing reflections.

Diffraction separation by variational mode decomposition

“…Seismic data are composed of dominant reflected data from large‐scale smooth reflectors, diffracted wavefields from small‐scale discontinuous points and background noise (e.g. Schwarz, 2019a; Lin et al ., 2020; Zhao et al ., 2020). Thus, equation () can be simplified by the decomposition of the coherent signals approximately composed of the reflected data and the diffracted wavefield: where R denotes the reflected data and D is the diffracted wavefield.…”

“…Thus, salt detection plays a vital role in oil and gas exploration and development. Seismic diffracted wavefields are the direct responses of these discontinuities and have the potential for high‐resolution imaging owing to their superior illumination, especially for real deepwater data (Schwarz and Gajewski, 2017; Lin et al ., 2020).…”

“…In addition, the diffracted events intersected with the reflection events several times at approximately 1 s, as indicated by the yellow circles. The conflicting‐dip events are a great challenge for traditional diffraction‐separation methods (Rad et al ., 2018; Lin et al ., 2020). Because of the presence of strong reflected data, the weak diffracted events are easily masked by the strong reflection events, and thus diffraction separation is necessary before diffraction imaging.…”

Prediction and Early Detection of Karsts—An Overview of Methods and Technologies for Safer Drilling in Carbonates

“…Imaging such small objects requires advanced acquisition and processing approaches. This problem has been investigated by [31,32].…”

Separation of diffracted waves via SVD filter

“…Shen et al (2019) also developed a method to enhance GPR diffracted waves using SVD filtering. Recently, Lin et al (2020) developed a method based on the multichannel singular-spectrum analysis algorithm to suppress time-linear signals (reflections) and separate weaker time-nonlinear signals (diffractions) in the common-offset or poststack domain. In fact, the SVD filter is a method based on the difference in the lateral coherence between different signals to achieve seismic wave field separation and denoising.…”