【摘要】 高频噪声压制是高分辨率地震数据处理中的关键性问题.基于广义S变换、经验模态分解的时频域地震去噪技术具有时变、分频和高保真特性,可有效处理非平稳地震信号,但二者在去噪原理、去噪效果、保真度、计算效率等方面尚存差异.对比分析表明:两种算法在提高地震信号信噪比的同时,可保持地震信号的保真度,保护陡倾角反射界面能量;基于广义S变换的叠前去噪方法以人工交互方式确定不同时刻的高频压制范围,以满足地震信号的时变和空变特性,但其计算效率较慢;基于经验模态分解的叠前去噪方法通过频率-空间域内的短窗口筛分出高频、高波数的本征模量函数,以消除高频、高波数噪声,但其高频噪声压制范围不能随时间变化,存在模态混叠的缺陷.更多还原

【Abstract】 It is a key issue that high frequency noise is suppressed in high resolution seismic data processing.The time-frequency domain denoising technique based on generalized S transform,empirical mode decomposition have some characters such as time-varying,frequency decomposition,high fidelity.Therefore they can effectively deal with the nonstationary signal.However,there are some differences between two algorithms in the denoising" principle, denoising effect,fidelity of signal and computational efficiency.It is shown that two algorithms not only can enhance the signal-noise ratio but also can effectively reserve the fidelity of seismic signal and protect the steeply dipping reflect event.The denoising method based on generalized S transform can reasonably determined the high frequency suppressed scope by interaction between man and computer in order to meet the need of time-varying and spatialvarying of seismic signal.But the computational efficiency of this method is relatively low.Moreover,the denoising method based on empirical mode decomposition can extract the high frequency and high wavenumber intrinsic mode function from the sliding time window in the frequency-spatial domain so it can restrain the high frequency noise.But the scope of high frequency noise suppression of this method can’t vary with time and there is a drawback of mode mixing in this method.更多还原