Ground roll suppression with synchrosqueezing wavelet transform in time-spatial domain

Ground roll is one of the main factors affecting the quality of seismic data. Suppressing ground roll is one of the vital steps in seismic data processing. Traditional ground roll suppression methods have inherent defects in time-frequency resolution and signal-to-noise ratio of processing results, which is not conducive to ground roll energy suppression. In this paper, the synchrosqueezing wavelet transform is extended to one-dimensional spatial domain, and a new method for suppressing ground roll by using synchrosqueezing wavelet transform in time domain and spatial domain is proposed. Firstly, the time-domain synchrosqueezing wavelet transform is used to separate the ground roll in the original seismic record. In order to reduce the damage to the low-frequency effective signal caused by the separation process, the first separated low-frequency data volume is subjected to the spatial synchrosqueezing wavelet transform to obtain the spatial wavenumber spectrum, determine the ground roll area, zero the coefficient in the area, suppress the ground roll and extract the effective signal, the extracted effective signal reconstruction results are superimposed on the separated ground roll seismic records to realize the joint suppression of ground roll. The synthetic and actual data processing results show that the combined method of time-domain synchrosqueezing wavelet transform and spatial-domain synchrosqueezing wavelet transform can effectively suppress the ground roll, greatly reduce the damage to the effective signal and improve the signal-to-noise ratio of seismic data.