Workflow structure

This workflow was created to enable the interpolation of high-resolution subbottom profiler data (2-6 kHz) into pseudo-3D cubes. Potential applications include:

• Pseudo-3D interpolation of:

  • subbottom profiler data (SBP)

  • single channel reflection seismic (SCS)

  • multichannel reflection seismic (MCS)

• Upsampling of existing post-stack 3D seismic volumes (e.g., in crossline direction)

The workflow can be subdivided into two separate stages (Figure 1):

1. 2D (pre-)processing of individual profiles
2. 3D interpolation of binned, sparse cube

Stage 1: Processing of SBP profiles (2D)

The most essential pre-processing steps for a subsequent 3D interpolation aim to correct the vertical offset between the input profiles. Without a sufficient correction, such offsets, especially at profile intersections, result in low quality interpolation outputs. This issue is addressed by three processing steps in the fist stage (Figure 1, upper panel):

• Static correction (per profile)
• Tidal heave compensation (per profile)
• Mistie correction (using all profiles)

2D stage: Input data types

For most of the scripts of the 2D processing stage, three different input types are allowed:

1. single SEG-Y file (e.g., filename.sgy)
2. datalist of files to process (e.g., datalist.txt)
3. directory with input files (e.g., /input_dir)

The datalist is a plain text file that contains the path and name of files to process relativ to the datalist.txt file:

./folder/file1.sgy
./folder/file2.sgy

for an examplary folder structure like this:

|-- folder/
|    |-- file1.sgy
|    |-- file2.sgy
|-- datalist.txt

Stage 2: Sparse data interpolation (3D)

In the second stage, the following processing steps are conducted (Figure 1, lower panel):

• Binning and geometry setup

• Preprocessing of sparse 3D cube

  • Gain application

  • Amplitude balancing

  • Frequency filter (lowpass, highpass, bandpass)

  • Trace resampling

  • Trace envelope

• Forward FFT (time to frequency domain)

• Interpolation via Projection Onto Convex Sets (POCS) algorithm

• Postprocessing of interpolated 3D cube

• Inverse FFT (frequency to time domain)

3D stage: Input data

Most scripts of the second, 3D processing and interpolation stage require a single netCDF file and, most often, one or more YAML configuration files.
An exception is the binning script, which needs multiple 2D netCDF files.

Figure 1: Schematic diagram of pseudo-3D interpolation workflow showing two stages: 2D processing of each profile (upper panel) and 3D interpolation of sparse 3D cube (lower panel).

---

Last update: Monday, 03 July 2023 at 09:46:51