The use of low-frequency seismic data improves the seismic resolution, and the imaging and inversion quality. Furthermore, low-frequency data are applied in hydrocarbon exploration; thus, we need to better use low-frequency data. In seismic wavelets, the loss of low-frequency data decreases the main lobe amplitude and increases the first side lobe amplitude and results in the periodic shocking attenuation of the secondary side lobe. The loss of low frequencies likely produces pseudo-events and the false appearance of higher resolution. We use models to examine the removal of low-frequency data in seismic data processing. The results suggest that the removal of low frequencies create distortions, especially for steep structures and thin layers. We also perform low-frequency expansion using compressed sensing and sparse constraints and develop the corresponding module. Finally, we apply the proposed method to real common image point gathers with good results.
