Accessible Recovery of Vintage Seismic Sections: From Paper to Migrated and Depth-Converted Data

In many regions, scanned images of vintage seismic reflection surveys constitute the only public subsurface information available, as the original data was frequently lost due to inadequate archival practices. Despite their age, these sections remain invaluable for geological studies; however, image files alone cannot be properly used in modern interpretation or processing software. To allow their scientific reuse, it is necessary to recover the original amplitude data through digitization: converting these images into the standard SEG-Y format. Current accessibility obstacles limit the widespread application of this process, as most existing tools are proprietary, unmaintained, or lack the precision required for advanced processing. Furthermore, digitization alone is insufficient; the full recovery workflow requires velocity model reconstruction and seismic post-stack processing to transform the digitized sections into data ready for interpretation.

To address these challenges, a new open-source digitization software and two complementary tools have been developed in Python: SEGYRecover, VelRecover and InSeis. SEGYRecover digitizes images of seismic sections by converting them into standard SEG-Y files. Built on established digitization methods, the program extracts the positive and negative amplitudes of individual traces plotted in variable-area wiggle displays and implements interpolation of clipped traces and smoothing of the final waveform. It also incorporates topography muting, trace mixing and gain correction to improve the visual consistency of the results. The resulting quality and fidelity of the digitized SEG-Y files are sufficient for advanced seismic analysis, including seismic attribute extraction, pseudo-relief generation, post-stack migration and integration with modern interpretation software. VelRecover enables the interpolation and visual editing of continuous velocity models from the sparse velocity analyses typically printed as headers alongside vintage seismic sections. These velocity models are suitable for depth conversion and post-stack migration. InSeis provides a graphical interface for building post-stack processing workflows using Seismic Unix through the Windows Subsystem for Linux, allowing users to apply deconvolution, post-stack migration, frequency filtering and other enhancements without command-line operations. The recovery of vintage seismic sections using this toolkit allows for the reinterpretation and detailed study of regions that was not possible with scanned sections alone.

Developed following FAIR principles and released under a GNU GPL v3.0 license, all three programs are freely available and fully documented via GitHub, Zenodo, and PyPI. This makes the digitization and processing of vintage seismic sections completely accessible to the global geoscience community.