Beamforming in Seismic Data Processing

AbstractMinimum variance unbiased (MVU) beamforming is a type of multichannel filtering which extracts coherent signals without distortion, whilst minimising residual noise power. Adaptive beamforming estimates signal and noise characteristics as part of the extraction process. This paper gives the fundamental principle of beamforming including signal and noise model, design criteria and the obtained beamformer. The adaptive beamformer used here is designed from models of primary and multiple reflection signals having parametrically specified moveout, amplitude and phase variation with offset (MVO, AVO and PVO). However, PVO is not usually justified in practice. The resulting analysis provides data for input into AVO and PVO schemes for obtaining lithological information. Synthetic data examples illustrate details of implementation of parametric adaptive MVU beamforming and the response characteristics of the resultant design. Real data examples show that data‐adaptive beamforming is more flexible and more effective in attenuating multiples in prestack common mid point seismic data than Radon transform methods. In common with other prestack multichannel processes, the advantages of beamforming are shown to best effect in data with a good signal‐to‐noise ratio.