Noise Reduction Of Desert Seismic Signals Using Adaptive Filtering Based On Spectral Kurtosis

With the exploitation of oil and natural gas,as non-renewable resources,the resources in shallow and easy-to-exploit areas are gradually consumed.Those unconventional reservoirs with complex environment and difficult exploitation have become a hot topic in recent years.Because of the difficulty of exploitation in desert area,there are still a large number of oil and gas resources that have not been exploited,and desert area is a very ideal oil and gas resource exploitation area.However,due to the low-frequency characteristics of desert noise and the large amount of spectrum overlap between noise and signal,the seismic exploration data obtained are extremely difficult to process,it’s unable to meet the "three high" requirements of high signal-to-noise ratio,high resolution and high fidelity,and it’s unable to enhance effective signals and reduce noise while maintaining high signal-to-noise ratio.Therefore,it is urgent to find a method of processing desert seismic exploration data in order to effectively exploit desert oil and gas resources.In traditional seismic exploration processing,TFPF(time-frequency peak filtering),wavelet transform and shearlet transform can achieve better denoising effect,and effectively process seismic data.However,when the above method is applied to desert seismic data processing,because of the complexity of desert seismic data,the processing effect is not obvious.Desert seismic data processing by band-pass filter has some effect,but it needs to find the frequency band artificially and can not solve the problem of signal and noise aliasing in the same frequency.To solve these problems,an adaptive spectral kurtosis bandpass filter based on spectral kurtosis is proposed to solve the problems in the traditional filtering methods.The filter first chooses the short-time Fourier transform(STFT)method to calculate the spectral kurtosis of noisy desert signals.Through the obtained spectral kurtosis,according to the spectral kurtosis,the energy transformation region of the signal is obtained,and the effective signal region and the noise signal region are divided.Reasonable threshold is selected,effective signal is retained and noise is discarded..By applying the processed spectral kurtosis value to the filtering algorithm designed in this paper,the effective signal after denoising can be obtained.Through the processing of the above method,the problem of searching frequency band artificially is solved.At the same time,compared with several traditional seismic exploration algorithms,it has better denoising effect,and the amplitude of the effective signal deserves better preservation.In this paper,the above algorithm is applied to the processing of simulated desert seismic noisy signal and actual desert seismic data.The results show that this method can suppress desert noise better,while retaining the effective signal better.Compared with several traditional seismic denoising methods,such as analog map,single channel contrast map,spectrum analysis map and actual record map,it can be seen that this method is more suitable for desert noise signal processing.The signal-to-noise ratio of noise suppression can be increased from-12.37 dB to6.09 dB.Through the above comparative experiments,it can be seen that this method is obviously superior to other traditional seismic denoising methods and solves the problem of desert noise signal processing.