The Study Of Noise Elimination Method Based On Curvelet Transform For Offshore Seismic Data

Seismic data processing is an important part of oil and gas exploration. For offshore seismic data acquisition, because of the special surface condion of offshore seismic data and influnced by external conditions, different levels of noise will be in seismic records, which destroy the continuity of effective wave and affect the quality of seismic profiles. To get the processing result with high signal to noise ratio and high definition, it is necessary for noise suppression.The main tasks of high-resolution exploration is to improve the signal to noise ratio of seismic data.Oil reservoir geophysics must have the data with high signal to ratio as a basis.The usual denoising method is only to focus on the improvement of the signal to noise ratio, of which fidelity analysis is not done well.How to improve the signal to noise ratio of seismic data as much as possible without causing serious distortion is the subject of noise suppression. The currently denoising methods used have advantages and disadvantages fidelity,it make sense that systematic study of their advantages and disadvantages,research for ways to optimize and improve methods and try to find new and effective methods to denoising.Curvelet transform is essentially derived from Ridge wave theory.It is a new multi-scale transformation developed on the basis of wavelet transform,in addition to the scale and location parameters,its structure elements also includes the orientation parameters which makes the Curvelet transform has good directional characteristics. Curvelet transform decomposition in all possible scales,it is a combination of a special filtering process and multi-scale ridge transform.Based on the multi-scale transform of curvelet,the sapce-time signal can be expressed rarefacted to obtain the optimal nonlinear approximation.By analyzing the seismic data characteristics in the curvelet domain,the different wave composition of space-time signals have differenc in curvelet domain,the separation of reflected wave and interference wave can be achieved from the frequence,angle and spatial location,at the same time the effective wave information can be better maintained.So the noise suppression based Curvelet has a strong practical significance.Because of its unique multi-scale and multi-directional,Curvelet has high recognition of position,If it is introducted into seismic data processin,the noise can be better removed,and the edge of denoised image can be maintained. The effective information can be retained while filtering out the noise.The method of seismic data denoising based on Curvelet can provide detailed information of more directions,which avoid the restrictions of other denoising methods to direction.On the basis of multi-scale decomposition in Curvelet domain and good approximation of the curve variation,this aritcle using threshold method to avoid the curve distorition caused by reconsturction.Form the processing point to consider the fidelity denoising method, this paper studies the basic principles of curvelet tansform and introduct it into the seismic data processing,ensured the relative amplitude of effective wave in the denoising process.Firstly,according to the noise calssification,the paper analysis the noise formation mechanism, understands the common noise characteristics ,such as energy, frequency, form and spatial distribution, get the basis of identification of various effective wave and established denoising method.Cuvelet transforms is applied to seismic data processing and through selection the appropriate threshold can get better denoising effect,and this method can provide more information for seismic data interpretation.