时频峰值滤波去噪技术及其应用

本文将时频峰值滤波(TFPF)去噪技术应用于共炮点地震资料的随机噪声压制.时频峰值滤波技术是通过频率调制将信号调制成解析信号的瞬时频率,利用解析信号的Wigner-Ville分布的峰值进行瞬时频率估计,恢复有效信号,与其它去噪方法相比,TFPF具有在较少的约束条件下压制强随机噪声的优点.本文针对实际地震资料的非线性特性,利用加窗的Wigner-Ville分布实现TFPF,使得地震信号在一个窗长内近似满足线性瞬时频率条件,减小由地震信号非线性引起的偏差.本文对共炮点地震记录做时频峰值滤波处理,滤波结果表明在地震勘探资料中存在强随机噪声的情况下,利用局部线性化处理的时频峰值滤波技术可以有效地压制地震资料中的随机噪声,恢复出湮没在随机噪声中的地震反射信号.信噪比提高3～6 dB.

Denoising technique based on Time-Frequency Peak Filtering and its application

The time-frequency peak filtering (TFPF) is applied to attenuating random noise in common shot point seismic record. TFPF is the technique that encodes the signal as instantaneous frequency (IF) of an analytic signal by frequency modulated. Then, the signal is recovered by getting peak of pseudo WVD of analytic signal according to frequency. Comparing with other denoising methods, TFPF has the superiority in using less constraint condition to attenuate random noise. For the nonlinearity of the seismic data, the WVD added window function is utilized to execute TFPF, which makes the IF in a window function be linear approximatively. The bias of filtered signal induced by nonlinearity can be reduced by local linearization. The TFPF with local linearization processing is applied to the common shot point data to attenuate random noise. The results show that the random noise can be attenuated well and the signal contaminated by strong random noise can be recovered after TFPF. The signal-to-noise ratio can be improved to 3~6 dB.