|
<< Click to Display Table of Contents >> Navigation: General > Minimum Phase Conversion for Vibroseis Data |
This module processes Vibroseis seismic data by cross-correlating the input gather with the pilot sweep to remove the sweep signature and convert the data toward minimum phase. For each input trace the module performs a frequency-domain correlation of the trace with the sweep signal using the provided sweep wavelet (or the first input trace if no sweep is connected), applies an optional band-pass taper to suppress noise outside the sweep frequency range, and optionally performs a deconvolution step to sharpen the output wavelet to minimum phase. The module outputs the processed gather, the vibroseis deconvolution filter operator, the estimated minimum-phase wavelet, the synthetic sweep, and the autocorrelation of the sweep.
Use this module as the primary correlation and minimum-phase conversion step in a Vibroseis processing workflow. Connect the output sweep from the Create Sweep module (or the recorded pilot sweep) to the Input sweep connection, and set the band-pass taper frequencies to match the sweep bandwidth.
Container for the standard input data connections. Connect this to the output of the preceding module in the processing sequence.
The raw, uncorrelated Vibroseis seismic gather to be processed. This is typically a shot gather or receiver gather containing the upcorrelated data before sweep removal.
Optional. The pilot or modelled sweep signal used as the correlation reference. If not connected, the first trace of the input gather is used as the sweep. Connect the output of the Create Sweep module or the recorded near-field sweep for best results.
When enabled, applies a spectral division deconvolution step after correlation to further sharpen the output wavelet toward minimum phase. When disabled (default: false), only the cross-correlation with the sweep is applied. Enable this option to improve temporal resolution at the cost of increased sensitivity to the noise level setting.
Container grouping the parameters that define how the correlation with the sweep is performed.
The length (s) of the cross-correlation output window. The default is 5 s. This should be set equal to the desired recording length of the correlated data. Typically this equals the listen time of the survey.
When enabled (default: true), a trapezoidal band-pass taper is applied to the correlation spectrum before the inverse FFT, limiting the output to the sweep frequency band. Disable this option only if you want to preserve energy outside the sweep bandwidth.
The low-cut frequency (Hz) of the band-pass taper applied during correlation. The default is 1 Hz. Set this to the low end of the sweep frequency range. Frequencies below this value are zeroed in the correlation spectrum.
The low-cut taper end frequency (Hz). The default is 5 Hz. Full amplitude is passed for frequencies between Fr2 and Fr3. Set this to the lower edge of the usable sweep bandwidth.
The high-cut taper start frequency (Hz). The default is 125 Hz. Set this to the upper edge of the usable sweep bandwidth, where the taper begins to roll off toward zero.
The high-cut frequency (Hz) of the band-pass taper. The default is 250 Hz. Frequencies above this value are zeroed. Set this to the maximum sweep frequency or the Nyquist frequency, whichever is lower.
The spectral whitening factor (fraction of peak amplitude) added to the sweep denominator to stabilise the deconvolution. The default is 0.01 (1%). Increase this value on noisy data to prevent spectral blow-up in low-energy frequency bands. Decrease it on high-quality sweep data to improve resolution.
Container grouping the parameters for the wavelet deconvolution step used to compute the minimum-phase filter and wavelet outputs.
Selects whether the minimum-phase deconvolution operator is estimated in the Frequency (default) or Time domain. The frequency-domain option is generally preferred as it is more stable. The time-domain option may be useful on very short gathers where FFT-based estimation is less reliable.
The noise regularisation (prewhitening) factor (fraction of peak power) added to the autocorrelation diagonal when computing the deconvolution operator using the Toeplitz solver. The default is 0.01 (1%). Increase this on noisy data to prevent instability in the shaping filter computation.
The length (s) of the deconvolution shaping filter operator. The default is 0.1 s (100 ms). Longer filter lengths capture more of the sweep's autocorrelation and allow the shaping filter to correct a broader range of spectral deviations from minimum phase.
Container for the standard output data connections including the correlated and minimum-phase converted output gather.
The correlated and optionally deconvolved output seismic gather. The sweep signature has been removed and the data has been converted toward minimum phase, ready for subsequent noise attenuation and deconvolution steps.
The shaping filter operator estimated from the sweep autocorrelation. This can be saved and applied to other datasets using the Matched filter Apply module.
The estimated minimum-phase wavelet derived from the sweep autocorrelation. Use this output for QC of the wavelet estimation and as input to wavelet-based deconvolution modules.
The sweep signal used as the correlation reference. Inspect this output to verify that the correct sweep was used in the correlation.
The autocorrelation of the sweep signal. Use this output for QC of the sweep quality. A sharp central peak with low side-lobes indicates a well-designed sweep with good autocorrelation properties.