GB1065870A - Seismic data processing - Google Patents

Abstract

1,065,870. Geophysical prospecting. TEXAS INSTRUMENTS Inc. June 18, 1964 [June 21 1963], No. 25308/64. Heading H4D. [Also in Division G1] In a method for minimizing reverberation effects in a seismic signal, particularly in a marine seismogram, the signal is processed by passing it through a delay line filter having an impulse time response which is approximately the inverse of the reverberation effect. In the embodiment shown in Fig. 2, seismic disturbances arising from the detonation of a shot 1 are detected by a linear array of seismometers whose outputs may be processed so as to provide a single noise-free signal f(t). The signal f(t) is fed to a delay line filter (5) (Fig. 3), which may be a magnetic drum type or a series of electrical delay circuits, having M + 1 outputs at equal time intervals. The signals from these outputs have their amplitudes multiplied by respective filter weighting factors # and are then combined in a summing network (6), such as a differential summing amplifier, whose output constitutes the processed signal. The required impulse time response of the filter (5) is obtained by adjustment of potentiometers (7). The values of the weighting factors are computed by recording the noise-free signal f(t) on magnetic tape and feeding it to an arrangement 3 which generates the autocorrelation function of f(t). Such an arrangement may comprise an analogue or digital computer using a sampling method to calculate a set of M + 1 auto-correlation function coefficients for the signal f(t). The weighting factors are uniquely defined in terms of the auto-correlation function coefficients by a matrix equation, and are calculated by a digital computer 4 using the auto-correlation function coefficients as input data. The power/frequency response of the delay line filter (5) is stated to be approximately inversely proportional to the power spectrum of the seismic signal f(t) and the weighting factors # may therefore be used to estimate the power spectrum of the signal f(t). Phase errors introduced into the seismic signal f(t) by the apparatus may be compensated for by the use of a second filter in front of the auto-correlation function generator 3. A detailed mathematical treatment of the autocorrelation principle, and numerical examples of auto-correlation function coefficients and their associated filter weighting factors #, are given. The mathematical relationship between the coefficients and their weighting factors is claimed.