Efficient seismic data acquisition with source separation
View Patent ↗A method for the simultaneous operation of multiple seismic vibrators using unique modified pseudorandom sweeps and recovery of the transmission path response from each vibrator is disclosed. The vibrator sweeps are derived from pseudorandom binary sequences modified to be weakly correlated over a time window of interest, spectrally shaped and amplitude level compressed. Cross-correlation with each pilot signal is used to perform an initial separation of the composite received signal data set. Recordings of the motion of each vibrator are also cross-correlated with each pilot, windowed, and transformed to form a source cross-spectral density matrix in the frequency domain useful for source signature removal and for additional crosstalk-suppression between the separated records. After source signature removal in the frequency domain an inverse transform is applied to produce an estimate of each source-to-receiver earth response in the time domain. The method has application to both land and marine geophysical exploration.
1. A method of seismic prospecting comprising the steps of:
selecting and generating a plurality of weakly correlated pseudorandom sequences;
modifying the weakly correlated pseudorandom sequences to minimize crosstalk;
generating source signals from the weakly correlated pseudorandom sequences; and,
transmitting the source signals generated from the weakly correlated pseudorandom sequences to the earth.
2. The method of claim 1 further comprising the steps of:
detecting and recording measured signals associated with each source;
detecting and recording a composite received signal in response to the transmission of source signal through the earth; and,
using the measured source signals to separate the transmission path response from the composite received signal.
3. The method of claim 1 wherein the step of generating a plurality of weakly correlated pseudorandom sequences further comprises a pseudorandom number generator.
4. The method of claim 3 wherein the pseudorandom number generator further comprises:
selecting a preferred pair of sequences; and,
combining the preferred pair of sequences using modulo-2 addition.
5. The method of claim 3 wherein the pseudorandom number generator further comprises:
selecting a preferred pair of sequences;
selecting a first parent sequence from the preferred pair of sequences;
selecting a second parent sequence from the preferred pair of sequences;
applying a delay to the second parent sequence;
combining the first parent sequence and delayed second parent sequence using modulo-2 addition to generate a Gold code.
6. The method of claim 3 wherein the pseudorandom number generator comprises a linear feedback shift register.
7. The method of claim 3 wherein the pseudorandom number generator comprises a pseudorandom number generator algorithm.
8. The method of claim 3 wherein the pseudorandom number generator comprises a Gaussian noise random number generator.
9. The method of claim 3 wherein the pseudorandom number generator comprises a hardware random number generator.
10. The method of claim 1 wherein the step of modifying the plurality of weakly correlated pseudorandom sequences to minimize crosstalk further comprises:
sampling the weakly correlated pseudorandom sequences;
generating a target spectrum;
reshaping the weakly correlated pseudorandom sequences by use of digital filtering;
computing the autocorrelations of the sequences and the cross-correlations between the various weakly correlated pseudorandom sequences;
windowing the autocorrelations of the sequences of the cross-correlations between the various weakly correlated pseudorandom sequences;
decoupling the weakly correlated pseudorandom sequences;
performing level compression of the weakly correlated pseudorandom sequences;
filtering the weakly correlated pseudorandom sequences for high frequency noise;
computing the autocorrelations of the sequences and the cross-correlations between the various weakly correlated pseudorandom sequences; and,
resampling the weakly correlated pseudorandom sequences.
11. The method of claim 10 , wherein the weakly correlated pseudorandom sequences are reshaped using the formula Target (x)=g1(x)g2(x) where g1(x) is (x/F1) 2 {[1−(x/F1) 2 ] 2 +[1.414*(x/F1)] 2 } −0.5 and g2(x) is 1 for x less than or equal to F2 and is {[1+cos ((x-F2)/(F3-F2)*π)]/2} 2 for x between F2 and F3, zero for x greater than F3, and F1=5, F2=100, and F3=150.
12. The method of claim 10 , wherein the weakly correlated pseudorandom sequences have bandwidths of about 10 Hz to about 100 Hz.
13. The method of claim 2 , wherein the step of separating the source signals from the composite received signal further comprising:
correlating measured source signals with source pilot signals;
correlating the composite received signal with the source pilot signal;
windowing the correlated data;
Fourier transforming the correlated data from the time domain to the frequency domain;
calculating the source separation matrix;
creating a receiver cross spectral density column vector for each frequency;
calculating the transfer function estimate by multiplying the receiver cross spectral density column vector by the source separation matrix; and,
inverse Fourier transforming the transfer function back to the time domain.
14. A method of operating a plurality of seismic vibrators comprising the steps of:
selecting and generating a plurality of weakly correlated pseudorandom sequences;
modifying the weakly correlated pseudorandom sequences to minimize crosstalk;
loading each vibrator with a unique weakly correlated pseudorandom sweep generated to minimize crosstalk;
activating all vibrators and using at least one detector to detect and record the composite received signal from all vibrators; and,
separating the composite received signal into individual transmission path estimates.
15. The method of claim 14 wherein the step of separating the composite received signal into individual transmission path estimates comprises:
correlating and windowing said composite received signal with pilot sweeps;
fast Fourier transforming the correlated and windowed composite received signals; and
applying a source separation matrix.
16. The method of claim 14 wherein the step of separating the composite received signal into individual transmission path estimates comprises correlating the composite recorded signal with the source pilot signals.
17. The method. of claim 1 . where the step of modifying the weakly correlated pseudorandom sequences to minimize crosstalk comprises decoupling a the weakly correlated pseudorandom sequences.
18. A method of operating a plurality of seismic vibrators comprising the steps of:
sampling a plurality of sweep sequences;
generating a target spectrum;
reshaping the sequences by use of digital filtering;
computing the autocorrelations of the sequences and the cross-correlations
between the various sequences:
decoupling the sequences;
performing level compression of the sequences:
filtering the sequences for high frequency noise;
computing the autocorrelations of the sequences and the cross-correlations between the various sequences;
resampling the sequences:
loading each vibrator with a unique one of said resampled sequences; and activating all vibrators and using. at least one detector to detect and record the composite received signal from all vibrators.
19. The method of claim 10 wherein the step of windowing the autocorrelations of the sequences of the cross-correlations between the various weakly correlated pseudorandom sequences employs a time window interval corresponding to approximately twice the listen time.
20. The method of claim 13 wherein the step of windowing the correlated data comprises windowing the correlated data over a time interval corresponding to approximately twice the listen time.