Large stores of seismic information currently exist on paper, film, and other media, containing unique records of earthquakes, geological exploration, and nuclear testing events. In order to become useful, the seismic data from the traces on these seismograms need to be extracted and outputted as computer‐readable digital data, a process referred to in this paper as digitization. Individual paper or microfilm seismograms can be digitized by hand and this tedious process has been ongoing since the 1950s. A few software packages were developed for semi‐automatic digitization, for example, the Teseo software for digitizing scanned seismogram images by Pintore et al. (2005). However, there is a pressing need to automate this process as much as possible, in order to realize the possibility of extracting the vast stores of information that reside in large archives worldwide. Although seismograms created since the 1980s are now readily available online in digital format, the need to digitize old seismograms is still necessary in order to apply modern techniques in studying pre‐1980 earthquakes (e.g., see Kanamori et al. , 2010; 2012) and nuclear explosions. We have made significant progress on this problem, working under a DOE SBIR Phase I grant, No. DE‐SC002717. The objective of our Phase I work was to create computer code that automatically extracts x – y digital data files of the seismic traces in seismogram image files, while addressing some of the problems associated with noise, distortions, and image discontinuities. In accomplishing this goal, we have developed a preliminary start‐to‐finish digitization process that begins with the inputted raster image, and ends with a digital data file representing trace center‐line coordinates. This process applies corrections to the images via our filters, and outputs, if desired, image overlays to allow for direct visual checking of the accuracy of the work. In this work, we used …