The new technology for suppression of hilly land seismic random noise: Shearlet transform and the high order weighted threshold function

As the seismic exploration environment is becoming more and more complicated, the SNR (Signal to Noise Ratio) of the obtained seismic data is much lower than before, the conventional method can not suppress the random noises effectively. Shearlet transform is a new multi-scale and multi-direction time frequency analysis method, the Shearlet transform has huge advantages in sparse representation characteristic and direction sensitivity, so Shearlet transform is suitable for seismic data processing. In conventional Shearlet denoising method, the hard threshold function is applied to choose the Shearlet coefficients. However, through hard threshold function, many valid signals are eliminated when the random noises are suppressed. This phenomenon leads to the appearance of false axis. In order to solve this problem, we propose the high order weighted threshold function, this new proposed threshold function has better continuity than hard threshold function and overcome the disadvantage of soft threshold function. Experiment shows the new method can eliminate the random noise of simulative seismic data and hilly land actual seismic data effectively and retain the amplitude of valid signals.