Page 14 - Acquisition and Processing of Marine Seismic Data

1. INTRODUCTION 5

SegD/SegY raw data
Data loading

Demultiplexing

Observer logs Geometry loading
Navigation logs

Band-pass filter
Near trace section
Gain recovery
Brute stack section
Track kill / Edit / Muting

f-k Dip filter

Multiple
suppression

Deconvolution
Band-pass filter

CDP sort / Binning
First velocity analysis
NMO
Prestack migration Velocity analysis
Dip moveout (DMO)
Inverse NMO
Gain NMO - Stack Second velocity analysis
NMO
Stack
Top mute Poststack migration
SEGY output Gain / top mute

FIG. 1.3 A conventional seismic data-processing flow for marine seismic data.

performed on the very first shots of the line in Fig. 1.4. Here, the dataset used has 500 shots
instead of applying the filter to the whole seis- with 480 recording channels, resulting in a total
mic line. After determination of all parameters of 240,000 traces and 120-fold seismic data. Sam-
for each step, the flow is run for the whole line pling rate and maximum recording time are
using approved parameters, which is known 1 ms and 6 s, respectively. Apart from the hard-
as production processing. ware specifications used to prepare this illustra-
Each of these steps is of a specific run time con- tion, the relative time span for each processing
trolled bydatavolume, processing hardware, and step is remarkable. In general, the processes per-
the algorithm implemented. As an example, nec- formed on prestack data require much more run
essary rendering times in seconds for different time thanthose applied to poststack data,because
steps for the same 2D seismic dataset are shown the data volume is significantly reduced after