A New Method to Suppress Random Seismic Noise Based on Curvelet Transform

This paper proposes a new method to suppress the random seismic noise by disposing the curvelet transform (CVT) coefficient. The curvelet coefficients of seismic signal after CVT still remain the characteristics of seismic event on different scales and along different directions, but the curvelet coefficients of noise distribute randomly. The curvelet coefficient, therefore, can be taken as the new target signal and de-noised. Then the processed CVT coefficients are used to restore the seismic data and to suppress the random seismic noises. Numerical experiment verified the correctness and validity of this method. Compared with the result obtained by discrete cosine transform and wavelet transform, the de-noising effect of CVT is better, and also better than that of only once CVT in signal to noise ratio.