Technology Research On The Crosswell Seismic Data Multi-Domain Abnormal Noise Suppression Method

Improving the resolution of seismic data is a key matter in the corsswell and groundseismic data processing process. While high resolution is based on the signal to noise ratioof good accuracy, because if the level of signal to noise ratio is low, the further treatment isno use. Due to the complex environment and seismic geological condition in the process ofcollecting seismic data, wave filed is quite rich, and includes many kinds of noises. But nofilter method is all-powerful, it cannot filter any noises, but only several special ones. So itis possible to filter noises in multi-domain in order to improve signal to noise by separatingsignificant wave and interference wave.It is an effective denoising method that transforming the seismic wave field fromtime-space domain to time-frequency domain and selecting proper filters to reducing noises.Comparing with the conventional spatial-temporal filtering or F-K filtering, it is muchbetter in denoising ability and fidelity, especially when the signal to noise ratio is relativelylow.In this paper, it analyses and deduces a wavelet threshold denoising, a waveletfrequency-devision damping filtering and weighted median filter based on the wavelettransform, aiming at the several noises of different natures in crosswell and surface seismicwave filed. The wavelet threshold denoising algorithm is also improved in denoising abilityand fidelity, and the testing of theoretical and practical corsswell and surface seismic dataindicates that it works well in reducing random, strong amplitude anomaly and lowfrequency noise.