一种有效的转换波去噪方法

One good denoising method of converted wave

为了提高转换波资料的信噪比,本文提出了利用反射波的振幅谱设计频率滤波器压制地震资料中的噪声方法.根据褶积理论反射波等于地震子波与反射系数的褶积,根据地震记录的假设地震记录等于反射波加上噪声.根据傅立叶变换公式对地震记录做傅立叶变换得到地震记录的傅立叶变换等于反射波的傅立叶变换加上噪声的傅立叶变换.使地震记录的傅立叶变换乘以反射波的傅立叶变换等于反射波的傅立叶变换乘以反射波的傅立叶变换加上噪声的傅立叶变换乘以反射波的傅立叶变换也等于反射波的傅立叶变换乘以反射波的傅立叶变换.对反射波的傅立叶变换乘以反射波的傅立叶变换做傅立叶反变换得到反射波地震记录.生成一个反射波地震记录,并计算该反射波的振幅谱.对该地震记录加上噪声,信噪比为100,利用计算的反射波振幅谱设计频率滤波器对信噪比为100的地震记录进行频率滤波.得到的滤波记录效果理想.对该地震记录加上噪声,信噪比为1,利用计算的反射波振幅谱设计频率滤波器对信噪比为1的地震记录进行频率滤波.得到的滤波记录效果也理想.利用常规带通滤波的方法对信噪比为1的地震记录进行频率滤波,滤波效果不理想.对实际的转换波资料进行滤波处理,利用带通滤波的方法计算不同频带的反射波的振幅谱,根据计算的振幅谱估计反射波的频率范围利用带通滤波方法对其进行频率滤波,获得带通滤波的处理结果,根据计算的振幅谱估计反射波的振幅谱利用本文方法对实际的转换波资料进行滤波处理,获得滤波的处理结果,对比带通滤波的结果和本文方法的滤波结果,本文的滤波效果理想.通过理论模型和实际转换波资料的噪声压制处理验证了当有效波的频率范围与噪声的频率范围不同时,带通滤波方法与本文方法处理效果相同,当有效波的频率范围与噪声的频率范围相同时,带通滤波方法的处理效果没有本文方法处理效果理想.

In order to improve the signal-to-noise ratio of converted wave data, this paper presents a method of using the amplitude spectrum of reflected wave to design frequency filter to suppress the noise in seismic data. According to the convolution theory, the reflection wave is equal to the convolution of seismic wavelet and reflection coefficient, and according to the assumption of seismic record, the seismic record is equal to the reflection wave plus noise. According to the Fourier transform formula, the Fourier transform of seismic record is equal to the Fourier transform of reflection wave plus the Fourier transform of noise. The Fourier transform of seismic record multiplied by the Fourier transform of reflection wave is equal to the Fourier transform of reflection wave multiplied by the Fourier transform of reflection wave. The Fourier transform of noise multiplied by the Fourier transform of reflection wave is also equal to the Fourier transform of reflection wave multiplied by the Fourier transform of reflection wave. The Fourier transform of reflection wave is multiplied by the Fourier transform of reflection wave, and the seismic record of reflection wave is obtained by inverse Fourier transform. A reflection seismic record is generated and the amplitude spectrum of the reflection is calculated. Adding noise to the seismic record, the signal-to-noise ratio is 100. Using the calculated amplitude spectrum of reflection wave, a frequency filter is designed to filter the seismic record with signal-to-noise ratio of 100. The result of filtering recording is ideal. When the signal-to-noise ratio is 1, the frequency filter is designed by using the calculated amplitude spectrum of reflection wave to filter the signal-to-noise ratio 1 seismic record. The filtering effect is also ideal. The conventional band-pass filtering method is used to filter the frequency of seismic records with signal-to-noise ratio of 1, and the filtering effect is not ideal. The actual converted wave data is filtered, and the amplitude spectrum of reflection wave in different frequency bands is calculated by using band-pass filtering method. The frequency range of reflection wave is estimated according to the calculated amplitude spectrum, and the frequency filtering method is used to obtain the processing result of band-pass filtering. The amplitude spectrum of reflection wave is estimated according to the calculated amplitude spectrum. The actual converted wave is analyzed by using this method Wave data filtering processing, filtering results obtained, compared with the results of band-pass filtering and filtering results of this method, the filtering effect of this paper is ideal. Through the noise suppression processing of the theoretical model and the actual converted wave data, it is verified that when the frequency range of the effective wave is different from that of the noise, the processing effect of the band-pass filtering method is the same as that of this method. When the frequency range of the effective wave is the same as that of the noise, the processing effect of the band-pass filtering method is not as good as that of this method.