Seistrak was a complete solution to the problem of seismic data that is only readily available as hardcopy plots. Seistrak works on scanned images of this hardcopy, accurately digitizes each trace and outputs the results in industry standard SEG-Y format – ready for workstation loading or further processing.

Seistrak is a complex software package, but the front-end user interface is intuitive and easy to understand, with on-line help and an integrated Quality Assurance system. The software works in stages and each stage is effectively quality controlled (on-screen) separately. This concept, of splitting down the complex process of image processing and data capture into separate steps, has tremendous advantages, not only for systematic quality control, but also for maximising productive throughput.

Making the most of paper seismic

Seistrak is a complete solution to the problem of seismic data that is only readily available as hardcopy plots. Seistrak works on scanned images of this hardcopy, accurately digitizes each trace and outputs the results in industry standard SEG-Y format – ready for workstation loading or further processing.

Seistrak is a complex software package, but the front-end user interface is intuitive and easy to understand, with on-line help and an integrated Quality Assurance system. The software works in stages and each stage is effectively quality controlled (on-screen) separately. This concept, of splitting down the complex process of image processing and data capture into separate steps, has tremendous advantages, not only for systematic quality control, but also for maximising productive throughput.

trak.gif (21585 bytes)

In summary the stages are:

1) Image alignment and distortion correction – this step effectively forces the scanned image into a rectangle and compensates for crooked paper alignment in the scanner and for any minor distortion of the original section (from paper stretch etc.) An unlimited number of control points can be entered on screen around the seismic plot by the user – the resultant shape is then transformed into a rectangle.

2) Timing Line Removal – a crucial process that helps the subsequent trace digitizing process. Seistrak actually removes the timing lines directly from the image file which can then be immediately browsed on-screen for effective quality control. The cleaned image can also be saved. The detection and removal parameters are totally under user control.

befcrop.gif (8398 bytes) aftcrop.gif (7383 bytes) 

3) Trace Position and Baseline Detection – in an ideal world, traces would be positioned vertically below the trace position detected at time zero. In practice, the trace position often drifts off the vertical. This may be due to minor paper stretch, or is more likely due to minor distortions on the image that arise when scanning a previously folded paper section. In some cases imperfections in the original plotter may also be a factor.

The error between the theoretical and actual trace position is usually less than one trace interval. Nevertheless this error can be enough to cause mis-tracking if it is not corrected for. Within Seistrak each baseline is tracked individually – the user can even quality control the baselines on screen.

4) Trace Digitizing and SEG-Y file output – Once all the above steps have been carried out (and the quality assurance flags have been set) the Seistrak software will then automatically digitize (or track) each trace, one after the other. First, tests are carried out by the user to optimise the tracking parameters and the results are quality controlled on screen. Various options are available, these include peak tracking together with clipped peak interpolation and trough tracking (a technique sometimes known as “full waveform” digitizing). If there are no troughs displayed then these can be interpolated. The final digitally sampled traces are written to a “raw” SEG-Y format disk file. The original CMP/SP numbers are written into the trace headers.

trakcrp.gif (4713 bytes)    linefin.gif (4967 bytes)

5) Basic Processing and Q.C.
Before being loaded to an interpretation workstation these “raw” files usually need to be filtered and amplitude balanced. As can be seen on the example opposite, the digitized traces (in green) show a very high frequency “jitter” due to the pixel structure of the image. In addition, any display bias on the original section will be apparent as a DC shift on the digital traces. Consequently, a simple bandpass filter (with frequency limits outside the original filters) is used to remove both these effects. This can be followed by a RMS trace balance. A QC plot is then produced for detailed comparison with the original.

6) Optional Further Processing
Once the original trace has been digitally captured, then significant improvements to the original section can usually be made by means of conventional post-stack processing routines. Particularly relevant for older vintage data are processes such as predictive deconvolution, FK or FX noise reduction filtering, migration, etc.

fillbit3.gif (6020 bytes)

 

Leave a Reply

*