Self-adaptive median strongnoise interference seismic data denoising method based on wavelet decomposition

The invention provides aself-adaptive median strong noise interference seismic data denoising method based on wavelet decomposition, which comprises the steps that seismic data are subjected to wavelet decomposition, a decomposition result is subjected to initial median filter processing; data is subjected to convolution treatment through a filter based on discrete cosine transform, the variance and kurtosis of obtained response data are estimated to further estimate the noise variance;based on a reference median filter window, the length of a filter window is reasonably selected according todifferent noise level estimations to define the length of a median filter window; the self-adaptive median filter is used for filtering the data, the length of the median filter window is adjusted ina self-adaptive mode according to thenoise levels of different regions; and the data subjected to median filtering treatment are reconstructed to obtain denoised seismic data. According to the self-adaptive median strong noise