Noise Attenuation for Seismic Signal by Non-Linear Thresholding in Curvelet Domain

Noise has traditionally been suppressed or eliminated in seismic data sets by the use of Fourier analysis filters and, to a lesser degree, nonlinear statistical filters. Although these methods are quite useful under specific conditions, they produce undesirable effects when denoising features of moderate to large amplitude and spatial extent, especially for the low S/N data. In this paper, a new method of de-noise in curvelet domain with non-linear thresholding is proposed: on the basis of curvelet multi-scale decomposition in good approximation of the curve variation characteristics, the author used non-linear threshold to address seismic data in curvelet domain. Through calculation of seismic data, we find out the method can suppress the random noise effectively while the effective wave can be maintained, the signal to noise ratio of the result is higher than the traditional method's. At the same time, it overcomes the drawback that the conventional filtering approach may affect the effective wave when suppressing noise.