Ground-roll noise attenuation by application of 2D discrete wavelet transform in radial trace domain

Summary: Reflection seismic data is often contaminated by a variety of coherent and incoherent noises that influence the reliability of the seismic data to provide a better understanding of the hydrocarbon reservoir characteristics. The groundroll noise is characterized by dispersive wave, low frequency, high amplitude that propagates along and near the surface of the earth and will often obscures the seismic reflection data. Removal of this type of noise is an essential part of land seismic data processing. Most of ground-roll noise removal or attenuation methods are based on transforms. Radial trace transform and wavelet transform are two of the most important and common transforms, which are used to attenuate ground-roll noise. Various methods have been introduced by different authors to attenuate the ground-roll noise. In this paper, we have used the wavelet transform on radial basis to attenuate the ground-roll noise. The method is based on the joint application of the two-dimensional discrete wavelet and radial transforms in order to eliminate the coherent noise, especially ground-roll noise. Since the radial trace transform reassign the linear events in the x-t domain into vertical events in the r-t domain, it provides a pleasant framework for two-dimensional (2D) discrete wavelet transform. The de-noised signal can be obtained by suppressing the vertical band of one-level, 2D discrete wavelet transform of the radial transformed seismic gather followed by transforming back the filtered radial-wavelet transform into the x-t domain. To evaluate the efficiency of the proposed method to eliminate the ground-roll noise, we have tested the method on a real reflection seismic gather from an oilfield in southwest of Iran. We have also compared the obtained results with those from the f – k filtering for ground-roll noise suppression from seismic data. The results obtained from applying the proposed method on the real data test prove that the proposed method for ground-roll noise attenuation is as good as, or better than, those produced using f –k filtering.
Ground-roll noise is the main type of coherent seismic noise and is characterized by dispersive wave, low frequency, high amplitude relative to other events of interest in land seismic surveys that propagates along and near the surface of the earth and obscures useful information seismic exploration. There has been much research about eliminating groundroll noise published in the literature, and many authors have introduced various methods for handling the ground-roll noise problem. A most straightforward and commonly used method for suppressing the ground-roll noise is band-pass filtering. In the case of the frequency overlap of primary reflections and ground-roll, the band-pass frequency filter either fail to attenuate all the ground-roll noise or remove much useful reflection energy. Dip filtering, also known as f-k filtering, which is based on the 2D Fourier transform, is another commonly used technique to attenuate ground-roll noise. In 2D Fourier transform of a seismic prestack gather, there is no a clear separation border between signal and noise region, consequently it leads to signal distortion. Another widely used method to remove ground-roll noise from seismic data is singular value decomposition (SVD)-based methods. Nevertheless, all of the above-mentioned approaches disagree with natural behavior of seismic data because of stationary seismic signal assumptions. Several methods have been introduced that consider the non-stationary nature for seismic data. The time-frequency de-noising algorithm is an effective method for handling noise problems. In this paper, we explain this method in detail, and analyze the influencing factors on the result of the method. Finally, the obtained results are compared with the results of the conventional f-k method.
Methodology and Approaches: We used a method, introduced by Almeida et al. (2015) based on the joint application of the radial and wavelet transforms in order to attain effective ground-roll noise attenuation. Since the radial trace transform reassign the linear events in the x-t domain into vertical events in the r-t domain, it provides a pleasant framework for 2D discrete wavelet transform. At the first, we perform the forward radial transform to remap of the seismic common shot gather to ensure efficient event separation by the wavelet transform. Next, the one-level, 2D discrete wavelet transform is applied on the output of the first stage. Due to the described inherent property of the four sub-bands of the 2D discrete wavelet transform and ground-roll characteristics, the ground-roll noise will appear mainly in the vertical (HL) band. The denoised signal can be obtained by suppressing the vertical band of one-level, 2D discrete wavelet transform and transform back the radial-wavelet transform into the x-t domain.
Results and The data set that we use to validate the efficiency of the proposed method is a common shot gather from demo data set of seismic data processing software VISTA. The obtained results show that, the proposed method is not sensitive to the type of mother wavelet and the geometric configuration of the origin of the radial trace transform. Hence, it can be an excellent alternative technique to ground-roll noise attenuation.