Study On Methods Of Noise Removing By Wavelet And Their Applications To Seismic Data Processing

Improving the signal-to-noise radio is an important task in the seismic dataprocessing, therefore the study of de-noising methods is always hot in seismicexploring. With the development of digital signals processing technology, manyexcellent de-noising methods come forth. In the seismic data de-noising, there is animportant research aspect that choosing the suitable de-noising method based on thefeature of seismic data.Wavelet analysis becomes a powerful tool in the field of nonstationary signalsprocessing in the last few years, because of its good ability of time-frequency analysis.In the same time, many de-noising methods based on wavelet transform are comingout. Furthermore they are proved to be superior to traditional Fourier transformde-noising methods. Four wavelet transform de-noising methods are studied in thispaper. The processing results indicate that the Mold maximum value wavelettransform de-noising has advantage in the low signal-to-noise ratio date, and themethod fits in the condition that contains white noise and many singularities. SpatiallySelective Noise Filtration is more fit in the of higher signal-to-noise ratio, and itscapability of edge reconstruction is better; Besides the WaveShrink method haswidespread applicability and its effect of de-noising is also very good; the translationinvariant wavelet threshold value de-noising method can restrain pseudo-Gibbs formWaveShrink threshold de-noising method, and get better Visual effect.The paper applies the four methods upwards to the seismic data processing,proves the validity of these methods through synthetic seismic recording and thepractical seismic data, then achieves the goal of improving the signal-to-noise radio inseismic data processing..