• Features
• System Requirements
• Installation
• Before You Begin
• Quick Start (5 Steps)
• Complete Tutorial
• Advanced Topics
• Citation
• References
• License

• Vectorization of seismic images - Convert paper/raster seismic sections into SEGY files compatible with modern interpretation software
• User-friendly GUI - Simple interface for the entire vectorization workflow
• Perspective correction - Handles skewed or distorted scanned images
• Automatic trace line detection - Identifies and vectorized individual traces, preserving the original number of traces
• Timeline detection and removal - Detects and removes horizontal timelines
• Amplitude extraction - Converts pixel density to the right of each trace to seismic amplitude values for each trace
• Frequency filtering - Apply bandpass filters to clean up vectorized data
• Geospatial referencing - Associates traces with real-world coordinates

• Operating System: Windows 10/11
• Memory: At least 8GB RAM
• Python: 3.12+ (automatically handled if installing via pip)
• Disk Space: Sufficient space for images and output files
• Image Requirements:
  • Binary images (black and white pixels only)
  • Supported formats: TIF, JPG, PNG
  • Variable area/wiggle display with positive amplitude filled in black

Windows Installation

1. Install Python (if not already installed):

   • Download Python from python.org
   • During installation, make sure to check "Add Python to PATH"
   • Click "Install Now" and wait for installation to complete

2. Install SEGYRecover:

   • Open Command Prompt (search for "cmd" in Windows search)
   • Type the following command and press Enter:

   pip install segyrecover
   

   Alternatively, install the latest development version directly from GitHub:

   python -m pip install git+https://github.com/a-pertuz/segyrecover.git
   

3. Launch the program: After installation, simply type:

   segyrecover
   

First Run Setup

When you run SEGYRecover for the first time:

• You'll be prompted to choose a data storage location
• Choose a location with plenty of disk space
• Example files will be copied to your selected location
• The application will create the necessary folder structure

Creating a Desktop Shortcut

1. Right-click on your desktop
2. Select "New" → "Shortcut"
3. Type segyrecover (if installed via pip)
4. Click "Next" and give the shortcut a name (e.g., "SEGYRecover")
5. Click "Finish"

File Organization

SEGYRecover uses the following folder structure:

segyrecover/
├── IMAGES/               # Store input seismic images
├── GEOMETRY/             # Store .geometry files with trace coordinates
├── LOG/                  # Store log files 
├── ROI/                  # Store region of interest points
├── PARAMETERS/           # Store processing parameters
└── SEGY/                 # Store output SEGY files

Prepare Your Data

1. Place seismic images in the IMAGES folder

2. Create geometry files (optional) in the GEOMETRY folder with format:

   CDP_NUMBER X_COORDINATE Y_COORDINATE
   100 500000.0 4500000.0
   101 500025.0 4500020.0
   

   • Files should have .geometry extension
   • Same base name as corresponding image files
   • Only first and last CDP points needed (software interpolates)

3. Verify image quality:

   • Higher resolution images yield better results
   • Minimal annotations overlapping seismic data
   • Clear trace lines and timelines

1. Launch → Run segyrecover and click "Start New Line"
2. Load → Select your seismic image file
3. Configure → Set parameters (trace coordinates, sample rate, frequency band)
4. Select → Mark three corners of your seismic section (ROI)
5. Process → Click "Start vectorization" and wait for completion

Your SEG-Y file will be saved in the SEGY folder and can be loaded into interpretation software.

Step 1: Loading an Image

1. From the Welcome screen, click "Start New Line"
2. Click "Load Image" to select your seismic section file
3. Navigate to your image and click "Open"
4. The image displays in the preview area with coordinate file info
5. Click "Next" to proceed to Parameters

Note: The console panel shows image details including dimensions and file path.

Step 2: Setting Parameters

Configure how the software interprets your seismic image:

1. After loading, you're taken to the Parameters tab
2. Previously saved parameters load automatically if available
3. Fill in the parameter sections below
4. Click "Save Parameters" to store settings
5. Click "Next" to proceed to ROI Selection

Region of Interest Corner Points

Map image pixels to seismic coordinates:

• P1 (Top Left): Trace number and TWT (Two-Way Time) value
• P2 (Top Right): Trace number and TWT value
• P3 (Bottom Left): Trace number and TWT value

Example:

• P1: Trace 100 at 0 ms (top left)
• P2: Trace 500 at 0 ms (top right)
• P3: Trace 100 at 3000 ms (bottom left)

Note: Use negative time values for data above datum (e.g., -200 ms)

Acquisition Parameters

• Sample Rate (ms): Time between data points (common: 2 or 4 ms)
• Frequency Band (Hz): Four-value bandpass filter:
  • F1: Start cutting below this frequency
  • F2: Keep everything above this frequency
  • F3: Keep everything below this frequency
  • F4: Start cutting above this frequency

Common Values:

• Vintage data: F1=8, F2=12, F3=60, F4=80
• Modern data: F1=3, F2=5, F3=80, F4=100

Detection Parameters

• TLT (Traceline Thickness): Vertical trace width in pixels (usually 1)
• HLT (Timeline Thickness): Horizontal timeline height in pixels (4-8)

Advanced Parameters

Default values work for most images

• HE (Horizontal Erode): Timeline removal aggressiveness (100+ px)
• BDB/BDE: Trace detection start/end rows from top
• BFT: Duplicate trace filtering strictness (0-100%, default 80%)

Step 3: Selecting Region of Interest

Mark three corners to define the vectorization area:

1. You're automatically taken to ROI Selection after parameters
2. Image displays with corner point selection prompt
3. Use magnifier 🔍 to zoom to top-left corner
4. Click "Top-left (1)" button, then click desired point

5. Use Home 🏠 to return to full view
6. Zoom to top-right corner, select (2) and click point
7. Repeat for bottom-left corner
8. Fourth corner calculates automatically
9. Verify result and click "Accept"
10. Click "Next" for vectorization

Note: ROI points save automatically for future reuse

Step 4: Processing

Automatic processing workflow:

1. Click "Start vectorization" to begin
2. Monitor console for step-by-step progress
3. Select CDP direction when prompted
4. Review processing in visualization window

Processing Steps

1. Timeline Detection and Removal

• Isolates horizontal timeline marks
• Removes timelines from original image

Timeline Detection Failure	Timeline Detection Success
HE parameter too low (200)	Correct isolation (HE=600)

2. Baseline Detection

• Identifies vertical trace lines
• Shows verification with green baseline overlay

3. Amplitude Extraction

• Counts black pixels per row between baselines
• Applies zero-value correction and smoothing
• Uses Akima interpolation for clipped values

4. Data Processing

• Resamples to specified sample rate
• Applies bandpass filtering

5. SEG-Y Creation

• Prompts for CDP direction (increasing/decreasing)
• Interpolates coordinates for all traces
• Creates standard SEG-Y with complete headers

Step 5: Results

View and analyze your vectorized data:

1. vectorized SEG-Y section displays in main view
2. Amplitude spectrum shows frequency content
3. Console provides vectorization summary
4. Additional processing options are available via buttons
5. Click "Start New Line" for next image

Additional Processing Options

The Results tab provides several buttons for further processing and editing of your vectorized SEG-Y data:

Edit SEGY Header

• Modify SEGY file headers including acquisition parameters and metadata
• Useful for adding extra information

Mute Topography

• Interactively define a muting surface by clicking on the seismic section
• Remove unwanted shallow data above a specified horizon
• Apply tapering to create smooth transitions
• Useful for removing noise above the topography, ground roll, or other unwanted noise

Apply AGC RMS

• Apply Automatic Gain Control using Root Mean Square method
• Balance trace amplitudes to enhance weaker signals
• Configurable gate length for amplitude averaging
• Helpful for improving visibility of deeper reflections

Apply Trace Mixing

• Enhance signal-to-noise ratio through trace averaging
• Choose between weighted average or median filtering methods
• Configurable mixing window size and weights
• Effective for attenuating random noise while preserving coherent signals

Each processing option opens a dedicated dialog with preview capabilities, allowing you to see the effects before applying changes. You can choose to save results as new files or overwrite the original.

Accessing SEG-Y Files

• File → Open Data Directory → navigate to SEGY folder
• File location shown in console panel
• Compatible with OpendTect, Petrel, Leapfrog, Kingdom

Troubleshooting

Poor Timeline Detection

• Increase HE (Horizontal Erode) parameter
• Ensure ROI aligns timelines horizontally

Missing/Extra Baselines

• Adjust TLT to match trace width
• Modify BDB/BDE for cleaner detection areas
• Increase BFT to filter false detections

Noisy/Spiky Data

• Adjust frequency filter parameters (F1-F4)
• Use narrower frequency band
• Check ROI excludes non-seismic elements

Coordinate Issues

• Verify geometry file format (CDP, X, Y)
• Match CDP numbers between geometry and parameters
• Use consistent coordinate system (UTM recommended)

Log Files

Check LOG folder for detailed process logs and error messages.

Frequently Asked Questions

How can I improve quality?

• Use highest resolution scans possible
• Experiment with frequency filter settings
• Ensure proper timeline detection/removal
• Make precise ROI selections

Batch processing available? Currently processes one image at a time. Use "Start New Line" sequentially and save parameters for similar images.

No geometry data? Still vectorizable but SEG-Y lacks real-world coordinates. Create placeholder file with first/last CDP numbers and arbitrary coordinates.

Color images supported? Works best with black/white images. Convert color seismic to grayscale/binary before processing.

Common Issues

• Program not found: Ensure Python added to PATH
• Missing dependencies: Run pip install <package_name>

If you use this software in your research, please cite it as:

Pertuz, A., Benito, M. I., Llanes, P., Suárez-González, P., & García-Martín, M. (2025a). SEGYRecover: A Python GUI-based tool for vectorizing vintage seismic reflection sections into SEG-Y files. Zenodo. https://doi.org/10.5281/zenodo.15053412

Find this software in the Zenodo Archive: https://doi.org/10.5281/zenodo.15053412

SEGYRecover uses several image processing and signal processing techniques. Some of them are covered by previous digitalization programs:

[1] Miles, P. R., Schaming, M., & Lovera, R. (2007). Resurrecting vintage paper seismic records. Marine Geophysical Researches, 28, 319-329.

[2] Farran, M. L. (2008). IMAGE2SEGY: Una aplicación informática para la conversión de imágenes de perfiles sísmicos a ficheros en formato SEGY. Geo-Temas, 10, 1215-1218.

[3] Sopher, D. (2018). Converting scanned images of seismic reflection data into SEG-Y format. Earth Science Informatics, 11(2), 241-255.

This software is licensed under the GNU General Public License v3.0 (GPL-3.0).

You may copy, distribute and modify the software as long as you track changes/dates in source files. Any modifications to or software including (via compiler) GPL-licensed code must also be made available under the GPL along with build & installation instructions.

For the full license text, see LICENSE or visit https://www.gnu.org/licenses/gpl-3.0.en.html