Automatic detection and attenuation of the external source interference noise by using a time-invariant hyperbolic Radon transform.

The external source interference noise (ESIN) in modern marine seismic data acquisition can seriously contaminate the reflection signal, which affects accurate reservoir characterization and description. An automatic ESINdetection and attenuation algorithm is proposed, which includes automatic ESIN detection, trajectory recognition and extraction. The ESIN detection aims to identify whether the shot gather is contaminated or not during the dynamic marine acquisition by using the average filtering and logistic operators. After obtaining the ESIN apex estimation and the optimal distance automatically, its trajectory can be recognized based on the derived time-invariant hyperbolic travel-time equation. For the recognized trajectories, a sparse time-invariant hyperbolic Radon transform (TIHRT) is designed in the mixed time-frequency domain to extract the ESIN by using the classification technique in the TIHRT domain. For implementing the TIHRT accurately and efficiently, an iterative 2D model shrinkage algorithm is introduced for the accurate and robust ESIN extraction and attenuation. Synthetic and field marine data examples are provided to demonstrate the promising performance of the proposed automatic ESIN detection and attenuation method. The proposed method can attenuate the detected ESIN effectively and preserve the reflection signals. [ABSTRACT FROM AUTHOR]