【摘要】 在野外地震数据采集过程中,如果地震测线上空有高压输电线通过时,在地震记录中就存在一个50HZ左右强单频千扰.这种干扰在地震记录的整个时间段上具有很强的能量,并且从浅层到深层频率、相位和振幅可能不相同,但是它们的变化不太大,可以认为基本保持不变.采用频率域压制方法很难完全消除这种干扰,仅仅是噪声和信号同时压制一定的倍数,因此并没有改善单频分量频率附近的信噪比.我们提出了基于余弦函数的自适应单频干扰消除方法,以有效消除地震数据单频干扰.在时间域内采用频率、时延和振幅的余弦函数来逼近这种单频干扰并从地震记录中减去,以实现消除这种单频干扰的目的.单频干扰余弦函数的振幅使用公式直接计算,单频干扰频率和时延采用自适应方法估算.本方法的最大优点就是能够有效消除干扰,而并不伤害干扰附近有效信号.因此在单频干扰频率分量附近数据信噪比提高了.通过实际数据的试算,说明了本方法的有效性和可行性.更多还原

【Abstract】 There is a strong mono-frequency interference around 50Hz in seismic records when a high-tension line crosses over a seismic line in the process of field seismic data acquisition. This interference has strong energy over whole time period of seismic records, and its frequency, phase and amplitude can be different from the shallow layer to the deep one. But their variations are not much too great, so they can be basically considered to maintain invariable. Adopting suppression method in frequency domain interference and signal have been only suppressed at certain amount simultaneously and the interference has hardly been removed, that is to say, signal-to-noise ratio has not been yet improved in the vicinity of mono-frequency component. So, a technique of the elimination of adaptive mono-frequency interference based on cosine function ( EAMFI ) is proposed to eliminate this interference in the seismic data. The mono-frequency interference is first expressed as a cosine function about frequency, time-lag, and amplitude in lime domain, then we estimate the frequency and time-lag of the mono-frequency interference using an adaptive method, and directly compute the amplitude of the mono-frequency interference using the formula of the amplitude computation, as well as the mono-frequency interference wave, finally subtract it from seismic record for the purpose of eliminating the interference wave. The most major advantage of this method is that only interference wave with mono-frequency component has been quickly and effectively eliminated, and useful signal has not been harmed, so the signal-to-noise ratio has been improved in the vicinity of mono-frequency component. We show that this method is feasible and effective through real data examples.更多还原