Reduction of seismic signal random noise based on grey filter

In the seismic signal processing and ground motion parameters obtaining, the seismic signal to noise ratio and the accuracy of ground motion parameters are mostly affected by the random noises. A novel method based on grey filter using GM (1, 1) model is proposed to solve this problem. When reducing Gaussian noises, the error of the grey filter are smaller than the median filter. In addition, the grey filter is effective when the distribution of the noises is unknown. The experimental results show that the grey filter is adapt to reduce Gaussian noises and distribution unknown noises. The grey filter model only needs four data so that it has advantage when the amount of seismic signal recorded data is small.