在地震资料中,噪声干扰严重影响了资料的品质,为此必须进行有效的去噪处理。通过对正交多项式变换方法的研究,本文提出一种基于奇异值分解(SVD)算法的正交多项式变换压制地震资料噪音的方法。将地震资料经过正交多项式变换后,有效信号和噪音在正交多项式系数上的分布不同,使用奇异值分解算法可以有效地分离有效信号和噪音。经合成地震记录模型及实际地震数据的处理结果表明,该方法不仅能够有效地进行信噪分离,而且可以减少对振幅的畸变,同时增强了反射同相轴的连续性。 In the seismic data, the noise interference seriously affects the quality of the data, for which must be effective denoising. Through the research on the orthogonal polynomial transformation method, this paper presents a new method of suppressing the seismic data noise by orthogonal polynomial transformation based on singular value decomposition (SVD) algorithm. After the seismic data is transformed by orthogonal polynomials, the distribution of the effective signal and the noise in orthogonal polynomial coefficients is different, and the effective signal and noise can be effectively separated by the singular value decomposition algorithm. The results of the synthetic seismogram model and the actual seismic data show that this method not only can effectively separate the signal and noise, but also can reduce the amplitude distortion and enhance the continuity of the reflection events.