Technique Research On Denoising Methods In2-D Crosswell Seismic

In various types of seismic data processing, the signal is alwaysaccompanied by a variety of noise. The goal of seismic data process is how toeffectively eliminate the noise, enhance signal-to-noise ratio by removing noise,to improve the resolution and image processing and lay the foundation forinterpretation.Polynomial fitting noise suppression technique is based on the premise ofthe seismic signal have a certain horizontal continuity, with the certainpolynomial to fit the seismic signal, every channel’s amplitude variations alongthe phase time can also be represented by fixed coefficient polynomial. First,calculate the phase time of seismic signal, the standard waveform and amplitudeweighting coefficients by polynomial fitting, and then they formed a fittingseismic trace, in order to enhance the signal continuity and improvesignal-to-noise ratio. Since the effective signal is continuous in space, thespecific approach is using multi-channel cross-correlation to determine theeffective wave event’s time and spatial location (ie, the signal time polynomial), then to determine the signal amplitude polynomial within the window by theleast square method, at last, obtain the standard waveform of significant wave inthis time window and repeat the above calculation steps in all time windowsdivided of the profile. So each window have calculated two coefficients of thepolynomial and a desired waveform. Multiply the desired waveform by thecalculated amplitude and put it on the position determined by the timepolynomial in order to finish the fitting on entire profile.For crosswell seismic and surface seismic data, based on the analysis andresearch of polynomial fitting denoising method principle, we compiled thepolynomial fitting denoising program. The following conclusions are obtainedby analysis and comparison utilizing theoretical models and practical processeddata.(1)Achieved random noise attenuation, effect of amplitude-preservation isquite good, also could separate wavefield and suppress the noise which apparentvelocity difference is large in a certain extent;(2)The test results show that, inthe program algorithm, the mixed wave coefficient and size of scanning windowis a major factor of the effect of noise suppression. Generally, small mixed wave coefficient or large window lead to a strong suppression, but wave field tends tostiff;(3)By comparing with the commercial software, the effect of thisprogram’s denoising process is obvious on both theoretical models and actualdata, the two are basically the same, so this program have a high practical value.